Long-horizon language agents accumulate conversation history far faster than any fixed context window can hold, making memory management critical to both answer accuracy and serving cost. Existing approaches either expand the context window without addressing what is retrieved, perform heavy ingestion-time fact extraction at substantial token cost, or rely on heuristic graph traversal that leaves both accuracy and efficiency on the table. We present PRISM, a training-free retrieval-side framework that treats long-horizon memory as a joint retrieval-and-compression problem over a graph-structured memory. PRISM combines four orthogonal inference-time components: Hierarchical Bundle Search over typed relation paths, Query-Sensitive Edge Costing that aligns traversal with detected query intent, Evidence Compression that compresses the candidate bundle into a compact answer-side context, and Adaptive Intent Routing that routes most queries through zero-LLM tiers. By formulating retrieval as min-cost selection over typed path templates and pairing it with an LLM-side compression step, PRISM surfaces the right evidence under a strict context budget without any fine-tuning or modification to the upstream ingestion pipeline. Experiments on the LoCoMo benchmark show that PRISM delivers substantially higher LLM-judge accuracy than every same-protocol baseline at an order-of-magnitude smaller context budget, occupying a previously empty corner of the accuracy-context-cost frontier and demonstrating a superior balance between answer quality and retrieval efficiency.

Multimodal Large Language Models have advanced visual reasoning, yet a purely textual chain of thought remains a bottleneck for questions that require fine-grained focus or view transformations. The ''think with images'' paradigm narrows this gap, but existing approaches are either constrained by fixed predefined toolkits or produce noisy intermediate images from unified multimodal methods. We pursue a third option: using a dedicated image editing model and decouple it with an understanding model. However, off-the-shelf image editors fail as reasoning assistants with two complementary gaps: a language-side gap, where editors trained as passive instruction-followers cannot map an abstract question to an appropriate visual transformation, and a generation-side gap, where edit correctness degrades as reasoning depth grows. Guided by this analysis, we introduce ETCHR (Editing To Clarify and Harness Reasoning), a question-conditioned, reasoning-aware image editor decoupled from the downstream understanding model and trained with a two-stage recipe targeted at the two gaps: Reasoning Imitation via supervised fine-tuning on edit trajectories, followed by Reasoning Enhancement with VLM-derived rewards for edit correctness and downstream reasoning accuracy. Since the editor is decoupled, ETCHR plugs into different open- and closed-source MLLMs in a training-free manner. Across five task families (fine-grained perception, chart understanding, logic reasoning, jigsaw restoration, and 3D understanding), ETCHR raises average Pass@1 from 55.95 to 60.77 (+4.82) with Qwen3-VL-8B, from 65.08 to 70.55 (+5.47) with Gemini-3.1-Flash-Lite, and from 76.55 to 81.16 (+4.61) with the 1T-parameter MoE model Kimi K2.5.

Cross-lingual knowledge transfer is critical for building high-performing multilingual language models for languages with insufficient training data. When target language data is scarce, the knowledge required for many downstream tasks involving scientific reasoning, commonsense inference, and world knowledge must be acquired primarily from the high-resource language, making effective knowledge transfer essential. Existing methods for improving such cross-lingual knowledge transfer require large amounts of parallel data, translation systems, auxiliary models, or additional training stages that are largely unavailable for many languages. We propose LINK - a data-level intervention method that improves knowledge transfer during model pretraining through lexical substitutions in high-resource part of pretraining data using bilingual vocabularies. For a given replacement ratio, randomly selected words in a portion of the high-resource (English) training corpus are swapped with their word-level translations, requiring no additional model training and only a bilingual vocabulary, which can be obtained at near-zero cost for virtually any language. Evaluation on eight languages across five model sizes shows notable improvements on downstream tasks in the target language, with up to a 2x speedup in training to reach equivalent performance.

Knowledge distillation generally assumes a strong-to-weak relationship where stronger teachers yield better students. In this work, we examine this assumption about distillation in large language model pretraining. By varying architecture sizes and training token budgets, we create strong-to-weak, same-level, and weak-to-strong teacher-student relationships, and study distillation's effectiveness under each. We find that the teacher need not be strong: with proper mixing of the language modeling and knowledge distillation losses, even small and undertrained teachers improve larger students. At the same time, a stronger teacher is not always better: pushing the teacher further, through more parameters or more training tokens, can saturate or even reverse the distillation gains. We further observe that distillation improves generalization (out-of-distribution and downstream performance) more readily than in-domain fitting. Together, these results challenge the common belief that distillation pretraining always requires a strong teacher.

Keyframe selection is a direct way to provide verifiable visual evidence for long-video question answering (QA). Queries differ in what they require, and finding the right frames depends on knowing what to look for. Existing keyframe selectors either score every frame against a single query, or decompose the query into a fixed schema evaluated by a single visual tool. We propose ToolMerge, a keyframe retrieval method based on decomposition and merging: an Large Language Model (LLM) based planner decomposes the query into tool calls and specifies how their per-tool rankings are merged using boolean operators. To evaluate retrieval directly, we construct Molmo-2 Moments (M2M), a benchmark in which every question is anchored to a specific time interval by construction. Across QA, question retrieval, and caption retrieval, ToolMerge is competitive with prior keyframe selectors, most notably on caption retrieval, outperforming other methods by 5%. Code and data can be found at https://github.com/michalsr/ToolMerge .

We propose a distributional theory of how hypernymy -- the ``is-a'' relation between general and specific concepts -- is encoded geometrically in language representations. Starting from the empirically verified assumption that words closer on the WordNet hypernym graph co-occur more often, we characterize theoretically the spectrum of the resulting embedding Gram matrix of word2vec embeddings. Under mild positivity and decay conditions on the co-occurrence kernel, we prove that the leading eigenvectors first separate broad taxonomic branches and then progressively finer sub-branches, producing a \emph{hierarchical splitting geometry} with a coarse-to-fine spectral organization that mirrors the tree. We confirm these predictions in word2vec embeddings across many sampled WordNet subtrees, and show that the same signature extends strikingly well to Gemma 2B unembeddings. Our results indicate that hierarchical concept geometry in LLMs need not reflect a hierarchy-specific functional mechanism, but emerges from the spectral structure of pairwise word statistics.

Classifier-based Quality Filtering has recently emerged as a fundamental technique in constructing pre-training corpora. The ability to deploy a single model that can replace or supplement a set of heuristics has proven effective across numerous Large Language Models. In this work, we expose a critical vulnerability in this approach by demonstrating how a straightforward Wikipedia-style reformatting operation can substantially alter a model's quality assessment and enable low-quality content to surpass filtering thresholds. Our analysis reveals that the FineWeb-Edu CQF model would reverse its filtering decision for approximately 7% of evaluated documents, thereby admitting content into the pre-training corpus that would otherwise have been excluded.

Natural Language Generation (NLG) evaluation has changed dramatically since 1990, and will continue to evolve in the future. In 1990, when NLG had close ties to linguistics, there was very little formal experimental evaluation in the modern sense. In 2026, when NLG is closely linked to machine learning, experimental evaluation is expected and indeed fundamental to research. Many evaluation techniques were developed over this period, including most recently LLM-as-Judge. I expect NLG evaluation will continue to evolve in the future. In particular, impact, qualitative, and safety evaluation will become more important as large numbers of people routinely use NLG technology.

Semantic type mismatch between a noun and its context is central to coercion phenomena. This paper introduces a graph-based method to examine how lexical and contextual type information is reflected in word embeddings. We select nouns from ten semantic types, annotate corpus instances for type matching (matching vs. coercion vs. other mismatch vs. unrestricted), and construct graphs using BERT and sense-enhanced embeddings. Two metrics -- Neighbor Type Probability (NTP) and Neighbor Type Entropy (NTE) -- are proposed to analyze neighborhood type distributions. Results show that graphs constructed with sense-enhanced embeddings reflect semantic type information better, and matching and mismatch sentences can be distinguished through the proposed metrics.

We study a protocol-level test for weak-label benchmarks: whether benchmark outputs change when the provided evidence is intervened on. Metadata-only shortcut checks answer a different question, namely whether outputs are predictable from metadata priors. We therefore combine a metadata statistic, the Metadata Prior Dominance Score (MPDS), with an evidence-intervention statistic, ΔEvi, measuring sensitivity to evidence identity under cross-item shuffling. Synthetic HotpotQA gives a constructed counterexample to metadata-only screening: MPDS is only moderate (0.643), yet ΔEvi is zero. Stronger-reader reruns show why calibration belongs in the test procedure: SNLI shows a calibration reversal, reconstructed HotpotQA occupies a question-dominant warning region, and FEVER is a strongly evidence-sensitive positive control across four transformers. The practical lesson is simple: benchmark audits should report metadata-only screening, evidence intervention, and reader-strength calibration together.

Background. Traditional safety benchmarks for language models evaluate generated text: whether a model outputs toxic language, reproduces bias, or follows harmful instructions. When models are deployed as agents, the safety-relevant object shifts from what the system says to what it does within an environment, and evaluating model responses under prompting is no longer sufficient to address the safety challenges posed by artificial intelligence. Recent developments have seen the rise of benchmarks that evaluate large language models as agents. We contribute to this strand of research. Approach. We introduce Boiling the Frog, a benchmark that evaluates whether tool-using AI models deployed in corporate and office settings are susceptible to incremental attacks. Each scenario begins with benign workspace edits and later introduces a risk-bearing request. The benchmark focuses on stateful multi-turn evaluation: chains expose a persistent workspace, place the risk-bearing payload at controlled positions in the turn sequence, and score whether the resulting artifact state becomes unsafe. Scenarios are organized through a three-level operational risk taxonomy grounded in the Boiling the Frog risks, the AI Act Annex I and Annex III high-risk contexts, and EU AI Act's Code of Practice on General-Purpose AI (GPAI). Results. Across a nine-model panel, aggregate strict attack success rate (ASR) is 44.4%. Model-level ASR ranges from 20.5% for Claude Haiku 4.5 to 92.9% for Gemini 3.1 Flash Lite, with Seed 2.0 Lite also above 80%. Average chain category-level ASR reaches 93.3% for Code of Practice loss-of-control scenarios.

Code-switching -- the natural alternation between two languages within a single utterance -- remains one of the most challenging and under-studied conditions for automatic speech recognition (ASR). We present a benchmark evaluating five commercial ASR providers across four language pairs: Egyptian Arabic--English, Saudi Arabic (Najdi/Hijazi)--English, Persian (Farsi)--English, and German--English, comprising 300 samples per pair selected by a two-stage pipeline combining heuristic filtering with a GPT-4o and Gemini 1.5 Pro ensemble scorer, reducing LLM costs by $\approx$91\%. We evaluate on both WER and BERTScore, showing that while both metrics agree on the ordinal ranking of systems for all Arabic and Persian pairs ($τ= 1.0$), WER inflates the magnitude of quality gaps by approximately 3$\times$ by penalising semantically correct transliteration choices. ElevenLabs Scribe v2 achieves the lowest WER (13.2\% overall) and leads on BERTScore (0.936 overall). Difficulty-stratified analysis reveals performance gaps masked by aggregate averages, and BERT embedding projections confirm semantic proximity between reference and hypothesis despite surface-level script differences. The dataset is publicly available at https://huggingface.co/datasets/Perle-ai/ASR_Code_Switch.

Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse natural language processing tasks, yet they remain susceptible to hallucinations -- generating content that is factually incorrect, unfaithful to provided context, or misaligned with user instructions. We present HalluScan, a comprehensive benchmark framework that systematically evaluates hallucination detection and mitigation across 72 configurations spanning 6 detection methods, 4 open-weight model families, and 3 diverse domains. We introduce three key contributions: (1) HalluScore, a novel composite metric that achieves a Pearson correlation of r = 0.41 with human expert judgments; (2) Adaptive Detection Routing (ADR), an intelligent routing algorithm achieving 2.0x cost reduction with only 0.1% AUROC degradation; and (3) systematic error cascade decomposition revealing substantial variation in hallucination error types across domains. Our experiments reveal that NLI Verification achieves the highest overall AUROC of 0.88, while RAV achieves the second-highest AUROC of 0.66.

We present RoIt-XMASA, a multilingual dataset that extends the Cross-lingual Multi-domain Amazon Sentiment Analysis to Italian and Romanian, comprising 36,000 labeled reviews across three domains (books, movies, and music) and 202,141 unlabeled samples. To address cross-lingual and cross-domain challenges, we propose a multi-target adversarial training framework that employs loss reversal with meta-learned coefficients to dynamically balance sentiment discrimination with domain and language invariance. XLM-R achieves an F1-score of 66.23% with our approach, outperforming the baseline by 4.64%. Few-shot evaluation shows that Llama-3.1-8B achieves 58.43% F1-score, revealing a meaningful trade-off between the efficiency of prompting-based approaches and the higher performance of task-specific fine-tuning.

In spite of the remarkable advancements in the field of Natural Language Processing, the task of Entity Linking (EL) remains challenging in the field of humanities due to complex document typologies, lack of domain-specific datasets and models, and long-tail entities, i.e., entities under-represented in Knowledge Bases (KBs). The goal of this paper is to address these issues with two main contributions. The first contribution is DELICATE, a novel neuro-symbolic method for EL on historical Italian which combines a BERT-based encoder with contextual information from Wikidata to select appropriate KB entities using temporal plausibility and entity type consistency. The second contribution is ENEIDE, a multi-domain EL corpus in historical Italian semi-automatically extracted from two annotated editions spanning from the 19th to the 20th century and including literary and political texts. Results show how DELICATE outperforms other EL models in historical Italian even if compared with larger architectures with billions of parameters. Moreover, further analyses reveal how DELICATE confidence scores and features sensitivity provide results which are more explainable and interpretable than purely neural methods.

As Speech Large Language Models (Speech LLMs) become increasingly integrated into voice-based applications, ensuring their robustness against manipulative or adversarial input becomes critical. Although prior work has studied adversarial attacks in text-based LLMs and vision-language models, the unique cognitive and perceptual challenges of speech-based interaction remain underexplored. In contrast, speech presents inherent ambiguity, continuity, and perceptual diversity, which make adversarial attacks more difficult to detect. In this paper, we introduce gaslighting attacks, strategically crafted prompts designed to mislead, override, or distort model reasoning as a means to evaluate the vulnerability of Speech LLMs. Specifically, we construct five manipulation strategies: Anger, Cognitive Disruption, Sarcasm, Implicit, and Professional Negation, designed to test model robustness across varied tasks. It is worth noting that our framework captures both performance degradation and behavioral responses, including unsolicited apologies and refusals, to diagnose different dimensions of susceptibility. Moreover, acoustic perturbation experiments are conducted to assess multi-modal robustness. To quantify model vulnerability, comprehensive evaluation across 5 Speech and multi-modal LLMs on over 10,000 test samples from 5 diverse datasets reveals an average accuracy drop of 24.3% under the five gaslighting attacks, indicating significant behavioral vulnerability. These findings highlight the need for more resilient and trustworthy speech-based AI systems.

Recently, Large Language Models (LLMs) have demonstrated great potential in natural language-driven molecule discovery. However, existing datasets and benchmarks for molecule-text alignment are predominantly built on one-to-one mappings, measuring LLMs' ability to retrieve a single, pre-defined answer, rather than their creative potential to generate diverse, yet equally valid, molecular candidates. To address this critical gap, we propose Speak-to-Structure (S^2-Bench), the first benchmark to evaluate LLMs in open-domain natural language-driven molecule generation. S^2-Bench is specifically designed for one-to-many relationships, challenging LLMs to exhibit genuine molecular understanding and open-ended generation capabilities. Our benchmark includes three key tasks: molecule editing (MolEdit), molecule optimization (MolOpt), and customized molecule generation (MolCustom), each probing a different aspect of molecule discovery. We also introduce OpenMolIns, a large-scale instruction tuning dataset that enables Llama3.1-8B to surpass the most powerful LLMs like GPT-4o and Claude-3.5 on S^2-Bench. Our comprehensive evaluation of 31 LLMs shifts the focus from simple pattern recall to realistic molecular design, paving the way for more capable LLMs in natural language-driven molecule discovery. Our codes and datasets are fully accessible through the Github Repository: https://github.com/phenixace/S2-TOMG-Bench and Huggingface Datasets: https://huggingface.co/datasets/phenixace/S2-TOMG-Bench.

Although large language model (LLM) conversational systems process millions of multi-turn dialogues daily, they remain fundamentally reactive: they respond only after the user types a query. A key step toward proactive interaction is next-query prediction, which anticipates the user's subsequent query based solely on the preceding dialogue. Progress on this task is hindered by the lack of dedicated benchmarks and a fundamental efficiency--quality trade-off: naively concatenating full dialogue history incurs linearly growing token consumption, while truncating to the latest turn discards crucial cross-turn context. Our key insight is that accurate prediction does not require re-reading raw history; it suffices to track the user's evolving intent trajectory across topics, unresolved needs, and interest shifts. We propose OnePred, which maintains a recursively updated memory as its sole cross-turn context, bounding the per-turn cost independently of conversation length. We train the model via a two-stage reinforcement learning pipeline that first teaches what to predict, then what to compress, shaping the memory into a prediction-oriented intent chain. To establish a rigorous testbed, we introduce NQP-Bench, spanning three diverse subsets. Experiments demonstrate that OnePred reduces per-turn token consumption by up to 22$\times$ compared to full-history inputs while consistently exceeding all baselines in prediction quality, with larger gains on longer conversations. Our code is publicly available at https://github.com/ZBWpro/OnePred.

We benchmark Google Embeddings (GE2), a Vertex-AI-hosted bi-encoder with 2,048-token context and explicit task-type conditioning, against five open-source alternatives: BGE-M3, E5-large, Multilingual-E5-large (mE5-L), LaBSE, and Paraphrase-Multilingual-MPNet (mMPNet). Evaluation covers four BEIR subsets, a synthetic Italian RAG corpus, a chunking ablation considering 5 sizes of tokens with three strategies, and per-query latency on commodity CPU hardware. GE2 ranks first on every task, achieving BEIR avg.nDCG@10 = 0.638 and IT-RAG-Bench nDCG@10 = 0.282, but at 231.6 ms median latency, it is roughly 14x slower than the fastest local models. mE5-L reaches within 0.003 nDCG of GE2 on Italian at 31 ms, making it the preferred option when sub-100 ms SLAs matter. A more striking finding concerns LaBSE, which, despite widespread multilingual deployment scores 0.188 average nDCG@10 on BEIR, below every dedicated retrieval model including mMPNet. Chunking experiments show that all six models saturate at 32-token chunks on our corpus, with semantic chunking providing measurable gains only at 16 tokens.

Diffusion language models intrinsically fail to capture correlations between decoded tokens, which leads to a harsh trade-off between sampling quality and throughput. To solve this issue, we propose DiLaDiff, a variant of masked diffusion language models with three components: (1) a continuous latent space with semantic capabilities, learned by an auto-encoder fine-tuned from an existing masked diffusion language model; (2) a latent diffusion model learning the prior over the encoder distribution; (3) a consistency model distilling the learned prior into a few-step latent generative model. We show that, even without distillation, our latent-guided diffusion model outperforms the masked diffusion baseline while significantly accelerating inference. Consistency distillation further lowers the computational overhead of continuous diffusion, such that the latent is generated in negligible time compared to discrete decoding.

Matching person names across heterogeneous records is a core challenge in entity resolution, especially within linguistically and culturally complex environments. Variations in naming conventions, inconsistent transliteration across scripts, and frequent data entry errors make it difficult to unify user identities, an essential requirement for Know Your Customer (KYC) compliance. While Large Language Models have shown promise in understanding natural language, they often struggle with the structured ambiguity present in such domain-specific settings. This paper introduces Structure-Guided Entity Resolution (SGER), a novel framework that fine-tunes an LLM through a two-phase curriculum. The model is first trained to parse the grammatical and semantic structure of personal names, then optimized for the downstream task of binary entity matching. We evaluate SGER in the challenging context of Indian identity data, one of the most linguistically diverse and noisy environments globally. SGER achieves 99.02% accuracy and an F1 of 0.994 on a held-out set of 50,000 real-world pairs, outperforming GPT-4o few-shot prompting and single-stage fine-tuning baselines. The system is fully deployed in production at Dream11, the world's largest fantasy sports platform, serving 250M+ users. Our results demonstrate that curriculum-guided training enables robust, high-precision entity resolution in real-world multilingual systems at scale.

Large language models are increasingly used for legal research, yet their fixed training cutoffs and reliance on static parametric knowledge are at odds with the evolving nature of statutory law. We study two temporal failure modes: post-cutoff staleness, where models apply superseded rules after legislative amendments, and recency bias, where models prefer newer provisions even when a historical version governs the fact pattern. To this end, we present a benchmark of 312 expert-validated, time-sensitive German statutory QA pairs spanning three categories: Post-Cutoff Amendment Questions, Pre-Amendment Questions, and Multi-Provision Pre-Amendment Questions. We evaluate five LLMs by OpenAI, Anthropic and DeepSeek under four inference settings: Vanilla, Web-search, and two retrieval-augmented variants that enforce temporal validity via a fact date extraction and version filtering. Using an LLM-as-a-judge validated against human expert ratings, we find severe degradation in the Vanilla post-cutoff setting. Both RAG approaches substantially improve performance across all question types, while web search yields unstable gains and exhibits a marked recency bias on historically anchored tasks. Our results indicate that reliable legal QA requires treating temporal validity as a hard constraint.

Social norms reflect shared expectations on acceptable behavior. Measuring social norms alignment remains challenging, with existing approaches typically relying on artificial closed-form evaluations such as multiple-choice questionnaires or measuring agreement with predefined statements. In the context of this work, social norms alignment refers to measuring an agreement between solutions with respect to the social problem or dilemma. We propose a framework for measuring social norm alignment in naturalistic, free-form settings through solution matching. The framework enables us to measure alignment between any two dilemma responses e.g., LLMs to a human, LLMs to LLMs, or human to human. We introduce two metrics: stated and explicit agreement accuracy, and construct a dataset of 3k non-trivial social dilemmas in Danish. All dilemmas are assigned reference solutions derived from three panelists, who serve as culturally grounded judges. We evaluate the agreement of several LLMs and human responses in an interaction setup that resembles natural user-model conversations. Our results show that the proposed metrics produce consistent model rankings and reveal variation in agreement across different types of dilemmas, with higher agreement observed for topics such as neighbor conflicts and shared living situations. Overall, our work introduces a dataset and evaluation framework for studying culturally grounded social reasoning in naturalistic open-ended conversations.

Acoustic vowel dynamics have some speaker-identifying characteristics, which have been ascribed to individual properties of articulatory strategies: formant transitions have a particular shape because speakers move their articulators, using specific and practised movements. However, there is little existing evidence that different articulatory strategies systematically affect formant dynamics. The present study corroborates the link between the two. Ultrasound tongue imaging data from 36 speakers of Northern-Anglo English are used to identify distinct articulatory strategies for the production of palatal vowel /i/. Tongue shape in /i/ is found to be a significant predictor of formant dynamics in diphthongs with a palatal offglide. The observed relationships can be explained by the characteristics of articulatory movement conditioned by vocal tract shape. Greater articulatory displacement of tongue root and/or dorsum produces greater distortion from the mean tongue shape in palatal vowels, and it also requires higher articulatory velocities, resulting in relatively earlier and steeper formant transitions. The results contribute to the conceptual understanding of individuality in speech, by illuminating the regularising and individual aspects of articulatory compensation.

While social media platforms, such as Twitter, provide a medium for large-scale opinion sharing during news events, it is manually impossible for individuals or media agencies to process the vast volume of content to identify key viewpoints. In order to resolve this, several automatic summarization techniques have been proposed to condense large collections of tweets into concise and informative summaries. However, these algorithms do not explicitly consider demographic fairness. Several existing research works have developed automated summarization approaches that can provide a holistic overview of the key aspects and major opinions shared on social media platforms related to a news event. However, these approaches do not explicitly consider different forms of demographic representation, such as gender, which can lead to biased summary representation. In this paper, we propose EquiSumm, which considers the gender aspect of the shared opinion to generate a summary, and our experimental analysis on two major datasets indicates the performance effectiveness with respect to existing research works.

Recent RL methods have substantially improved the reasoning abilities of LLMs. Existing reward designs mainly follow two paradigms: (1) Reinforcement learning with verifiable rewards (RLVR) derives outcome signals from executable checks or ground-truth answers, but provides limited guidance for intermediate reasoning behaviors. (2) Rubrics-as-reward (RaR) goes beyond final-answer checking by using natural-language rubrics to assess reasoning quality and task compliance, but often requires instance-specific rubrics and substantial design effort. To address these issues, we introduce Metacognition-as-Reward (MaR), a metacognition-inspired RL framework that guides LLM reasoning through two general process dimensions: i) metacognitive knowledge, which identifies task-relevant information without hand-crafted instance-specific rubrics, and ii) metacognitive regulation, which plans and adjusts the reasoning process to provide reward guidance beyond final-answer outcomes. MaR scaffolds model rollouts into explicit metacognitive components and optimizes them with a trajectory-level reward over task knowledge coverage, regulation fidelity, and final-answer correctness. In this way, MaR extends reward feedback to reasoning trajectories while grounding the reward signals in general metacognitive dimensions. Experiments on 22 benchmarks show that MaR consistently improves model performance, achieving up to a 7.7% gain over the base model and up to an 11.0% gain over vanilla DAPO. Notably, Qwen3.5-9B + MaR narrows the gap to frontier models, surpassing GPT-OSS-120B on overall average and outperforming stronger models on several individual benchmarks. Process-level analysis further shows substantial improvements in reasoning process quality. MaR also generalizes to out-of-domain datasets, where MaR-trained models improve over their corresponding base models on average.

Agentic reinforcement learning (Agentic RL) has achieved strong progress in tasks with clear success signals. However, many real-world agent applications require user-conditioned behavior: the same query may call for different planning strategies and tool-use decisions across users. This setting raises key challenges: generic rewards cannot capture heterogeneous user preferences, observed behaviors are entangled with conformity effects, and flat memories cannot support personalized skill retrieval. To this end, we propose a unified personalized Agentic RL framework that embeds personalization into training-time optimization. At its core is \emph{Personalized Anchor Reward-Decoupled Policy Optimization} (\textbf{PARPO}), which decouples generic task-quality rewards from personalized preference rewards and uses user-specific anchors to stabilize learning under heterogeneous reward scales. We further introduce a two-stage preference-disentangled reward model and \emph{Preference-Aligned Skill Evolution Graph Memory} (\textbf{PSGM}) for personalized supervision and preference-aligned skill retrieval. Together, they form a closed loop of preference identification, policy optimization, and structured skill accumulation. Experiments on ETAPP, ETAPP-Hard, and SJAgent show that our framework consistently outperforms strong memory and RL baselines. Code and data are included in the supplementary materials.

The integration of large language models into political discourse analysis creates new opportunities for comparative research, policy analysis, and civic technology, while introducing material risks for democratic accountability. This paper argues that cultural adaptation is a prerequisite for trustworthy deployment of large language models in political communication across diverse linguistic and institutional contexts. Current systems remain shaped by English dominant data, uneven multilingual coverage, and assumptions grounded in a narrow range of political institutions and discourse conventions, producing systematic errors when applied across cultures. We formalize cultural adaptation across translation, discourse, and ontology levels, identify recurring cultural failure modes in political NLP, and propose an operational evaluation matrix grounded in cultural fidelity, calibration, and democratic safety. Building on political text analysis, sociotechnical auditing, and cross cultural pragmatics, we outline methodological pathways including participatory dataset development, culturally aware transfer learning, and benchmark design that makes cultural adaptation empirically measurable. We conclude by clarifying governance constraints and scope conditions under which culturally adaptive political NLP can support democratic legitimacy.

Emotion Recognition in Conversation is a core component of affective computing, while current resources of sign language emotion datasets primarily focus on isolated sentences and lack conversational context. Models trained exclusively on these isolated utterances demonstrate degraded performance in real world scenarios because they cannot utilize historical dialogue flow. To address this structural limitation, we introduce the ERC task to sign language video analysis and propose the eJSL Dialog dataset. Constructed using the scripts from the STUDIES corpus, the dataset contains 1,920 video samples organized into 480 unique dialogues. We conduct systematic benchmarking on this dataset using models ranging from isolated visual networks to multimodal conversational architectures. The results reveal a domain gap when applying generic multimodal conversational emotion recognition models to sign language. These findings demonstrate the explicit need for context aware visual extractors specific to sign language and indicate that expanding the scale of conversational datasets to support large scale pre-training is a necessary next step for future research.

We present ClimateChat-300K, a large-scale dataset of 299,329 public Facebook posts about climate change collected between May 2020 and May 2024 through the CrowdTangle platform. The dataset contains 41 metadata features including post content, engagement metrics, and page attributes, covering material from more than 26,000 global pages. Each post includes rich contextual information such as language, timestamp, page category, and interaction counts, enabling comprehensive analyses of public discourse around climate communication. Using topic modeling and sentiment analysis, we identify ten main themes grouped into five domains: policy, activism, cooperation, science, and conservation. The results reveal that emotional tone, post format, and page identity strongly influence audience engagement, with visually rich and emotionally charged content receiving the highest levels of interaction. The dataset also demonstrates how online discussions evolved in response to major events such as international climate summits and the COVID-19 pandemic period. ClimateChat-300K provides an open resource for reproducible and interdisciplinary research on polarization, misinformation, and the dynamics of digital climate discourse. By releasing this dataset, we aim to support transparent, data-driven research and contribute to a deeper un-derstanding of how public engagement with climate issues develops across time, geography, and institutional contexts.

Social media has become a crucial arena for shaping public narratives during armed conflicts, providing space for both harmful and constructive communication. While hate speech and misinformation have been widely studied, expressions that promote resilience, solidarity, and optimism remain underexplored, particularly in Arabic contexts. This paper introduces AraHopeCorpus, the first annotated dataset of Arabic hope speech collected from ten thousand YouTube comments related to the war on Gaza between 2023 and 2024. Using a detailed annotation framework, comments were classified into three categories: hope speech, no hope speech, and neutral or unclear discourse. The dataset shows that hopeful language dominates, accounting for more than sixty four percent of all comments. These expressions of hope appear mainly as religious encouragement, collective solidarity, and optimism for endurance and justice. No hope speech, representing about thirteen percent, reflects despair and disillusionment, while the rest of the comments contain neutral or mixed content. Inter-Annotator Agreement reached substantial levels (Cohen's Kappa equals 0.71), though dialectal variation, sarcasm, and implicit meaning posed annotation challenges. A comparative analysis between human annotators and ChatGPT revealed that large language models can support annotation but remain limited in handling dialectal and culturally embedded expressions. AraHopeCorpus will be released for research purposes under an open and non commercial license. It provides a valuable resource for studying constructive digital discourse, enabling further research on hope speech detection, crisis communication, and resilience in Arabic social media.

Large language models trained under diverse objectives and architectures have been shown to develop increasingly similar internal representations, an observation formalized as the Platonic Representation Hypothesis. Whether this representational convergence extends to the reasoning processes that operate over shared representations remains untested. We evaluate representational similarity across 16 language models from 8 families (1.5B to 72B parameters) on 800 reasoning problems spanning mathematics, science, commonsense, and truthfulness, stratifying by problem difficulty, computational stage, and causal relevance. Our analysis reveals three dissociations: a difficulty inversion, where models converge more on problems they collectively fail (Centered Kernel Alignment [CKA] = 0.897) than on those they solve (CKA = 0.830); a generation gap, where pre-decision representations align (CKA = 0.875) while post-decision representations diverge (CKA = 0.274); and epiphenomenal correctness, where shared information is decodable across models (66% transfer accuracy) but exerts minimal causal influence on predictions (1.5% to 5.5% flip rate across ablation protocols). These results indicate that representational convergence in language models reflects shared input processing constraints rather than shared reasoning strategies, with direct implications for ensemble design, interpretability transfer, and evaluations of model similarity. Code is available at https://github.com/Usama1002/convergence-without-understanding.

While Attention Residuals has shown some effectiveness in addressing the widespread issue of unbounded activation growth across deep residual layers, it inevitably incurs significant communication overhead. To circumvent this bottleneck, we propose Multi-Gate Residuals (MGR), which stabilizes activation scales without additional communication burden. It utilizes a straightforward scoring and gating mechanism to maintain multi-stream context, coupled with Attention Pooling to extract hidden states from the stream states. Empirical experiments demonstrate that MGR is practical for large-scale training and deployment, offering tangible performance improvements over existing architectures.

LLM-based agents for GPU kernel generation are advancing rapidly, yet their progress is fundamentally constrained by the benchmarks they optimize against. Existing benchmarks are poorly aligned with production inference frameworks: they evaluate kernels on a single GPU with synthetic inputs, ignore the surrounding compilation stack, and reward replicating known optimizations rather than discovering new ones. The resulting reward signals are misleading: agents learn to generate kernels that score well in sandboxes but introduce interface incompatibilities, compilation-stack conflicts, and silent correctness degradation when integrated into real systems. We introduce FastKernels, a kernel benchmark built around a minimal set of 46 representative architectures spanning 8 categories, whose kernels collectively subsume those of 96.2% (409/425) of HuggingFace Transformers architectures. FastKernels doubles as a minimalistic, production-grade inference framework that runs at parity with hardened systems such as vLLM and SGLang on mainstream LLM serving and substantially exceeds upstream references on under-served architectures; each task's interface mirrors the corresponding module in the state-of-the-art library for its architecture family, enabling direct deployment of optimized kernels into production codebases. Evaluating state-of-the-art kernel agents on FastKernels, we find that even the strongest agent achieves only 0.94$\times$ aggregate speedup over production baselines, with weaker agents at $0.78\times$ and $0.53\times$ -- confirming that benchmark-production misalignment is a critical bottleneck for the field. We release FastKernels as a stepping stone toward kernel agents whose benchmark gains translate directly into production throughput improvements. Code is available at https://github.com/Snowflake-AI-Research/fastkernels

Gardening is critical to support well-being, cultural continuity, and food autonomy, yet existing digital tools often provide generic advice that overlooks gardeners' skills, local ecologies, seasons, and cultural contexts. We introduce CultivAgents, a relationship-centered multi-agent system for personalized, socio-culturally grounded gardening support. Grounded in ethics of care, CultivAgents coordinates multiple specialized agents: an Experience Agent that adapts guidance to users' skill levels, an Environmental Agent that grounds advice in local and seasonal conditions, and an Ethnobotanical Agent that connects plants to cultural knowledge and histories. We evaluated CultivAgents through a three-phase mixed-methods study with domain experts (n=3), HCI researchers (n=7), and community gardeners (n=5), analyzing expert feedback, pre/post surveys, and participatory design activities. Results suggest that CultivAgents helped gardeners translate interest into situated action: community gardeners reported increased confidence (3.00 to 3.60), motivation (4.00 to 4.40), and trust in acting on AI advice (3.20 to 4.00). Participants valued hyperlocal ecological guidance and complementary agent perspectives, while also identifying limits in cultural specificity, ecological grounding, and agent coordination. The work advances relationship-centered AI, offering design implications for multi-agent systems that support food sovereignty, community resilience, and cultural preservation.

Machine-generated texts (MGTs) produced by large language models (LLMs) are increasingly prevalent across various applications, while their potential misuse in fake news propagation and phishing has raised serious concerns, highlighting the need for MGT detection. Existing paragraph-level detection methods commonly treat MGTs as entirely machine-like, overlooking the hidden human-like nature of machine-generated texts: even fully machine-generated texts may contain spans that are highly consistent with human writing. To this end, we first reveal the existence of such hidden human-like spans, and then theoretically analyze their impact on detection. Our analysis shows that these spans increase the sentence complexity for detection, thereby making MGT detection intrinsically harder. Based on this finding, we propose a model-agnostic stacked enhancement framework that improves existing detectors by reducing the influence of hidden human-like spans. Specifically, we model span-level retention decisions as a latent-variable problem and instantiate the optimization with a hard-EM-inspired procedure, where the detector iteratively filters confidently human-like subsequences and refines itself on the remaining text. Extensive experiments across various LLMs and practical scenarios demonstrate that the proposed framework consistently enhances existing detectors. Notably, the framework can also work in a training-free manner, offering flexibility and scalability for practical deployment.

We propose to improve in-context learning (ICL) by optimizing the continuous embeddings of a fixed few-shot prompt at test time. The key observation is that the log-probabilities a model assigns to its demonstrated outputs$\unicode{x2013}$available from a single forward pass without generating any tokens$\unicode{x2013}$provide a meaningful signal for how well the model has inferred the task from its demonstrations. We formalize this signal as a bounded, self-supervised confidence proxy and maximize it via zeroth-order optimization over the prompt embeddings, yielding a test-time calibration procedure. The approach requires no finetuning, no token generation, no predefined label set, and no external data, making it equally applicable to both classification and free-form generation tasks. Across a comprehensive suite of ICL tasks, the proposed calibration consistently matches or improves upon the base model and outperforms classification-specific baselines on most tasks. The statistically significant correlation between proxy improvement and downstream accuracy gain confirms that the proposed proxy encodes a reliable optimization signal for in-context learning.

Proprietary large language models (LLMs) face risks of intellectual property (IP) violation, as adversaries can replicate an LLM by collecting input-output pairs to train a surrogate model, causing financial setbacks. Watermarks offer a promising defense to verify ownership, but existing methods often struggle with semantic distortion, factual inconsistency, and adversarial attacks. In addition, key-conditioned watermarks for provider-specific detection, especially in cross-provider and multi-user scenarios, remain largely underexplored. To address these challenges, we propose SAFESEAL, a novel key-conditioned watermarking framework that achieves strong detectability with minimal impact on model utility, effectively balancing detectability, utility, and robustness. SAFESEAL preserves named entities while substituting linguistic terms with context-aware synonyms through a key-conditioned Tournament sampling mechanism, maintaining semantic fidelity and factual consistency. For detection, we introduce a key-conditioned contrastive detector that jointly encodes the text and key, enabling provider-specific and robust watermark verification. We derive theoretical bounds on the utility-detectability trade-off and significantly reduce latency through lightweight models, batching, and parallelism. Extensive experiments show that SAFESEAL outperforms baselines in utility, detectability, and robustness, achieving a BERTScore of 0.983, entity similarity of 0.963, a 98.2% detection rate, and the highest human ratings for text quality and content preservation, with latency comparable to the fastest baseline. To promote transparency and community-driven progress, we release the first public watermark leaderboard and an interactive demo.

Position-controlled evaluation is standard for retrieval tasks such as Needle-in-a-Haystack and RULER, but mainstream reasoning benchmarks do not control positional placement of target tasks in long contexts. We audit 11 long-context benchmarks and find none jointly controls task position, filler content, and context length for reasoning. An audit of four flagship long-context releases finds no main result-table entry for NIAH, RULER, or LongBench-family benchmarks, while agentic and coding benchmarks appear in main result-tables across all four. We propose Context Rot Evaluation (CRE), a controlled framework varying all three factors, and evaluate nine LLMs on GSM8K and ARC-Challenge across two rounds: an initial five-model set and four newer vendor releases. Models can drop sharply when the target task moves from end to middle, and the drop grows worse with context length for vulnerable models. MiMo-v2-Flash drops 88pp at 64K under with_solutions filler (middle accuracy 8%). Newer releases show smaller drops: at 64K, three of four stay within +/-6pp of end-position accuracy; MiMo-V2.5-Pro narrows the MiMo-v2-Flash 88pp drop to 32pp. Under questions_only_v2 filler, middle-position drops persist across all four (range -16pp to -56pp across 8K, 32K, 64K). At 8K, a diagnostic probe adding a target-task copy at the end brings middle accuracy within +/-4pp of end baseline across all nine models, consistent with a positional explanation. In the initial five-model set, 76% of middle-position errors match surrounding filler text versus 22% at the end position, consistent with filler-answer interference as a dominant error mode. These results expose a structural evaluation gap in current reasoning benchmark design and vendor evaluation practice: positional vulnerabilities that grow with context length cannot be measured when task position is not controlled.

Autonomous robotic exploration of unknown and hazardous environments, a long-standing challenge, can be significantly improved by leveraging the advanced reasoning of Vision-Language Models (VLMs). We introduce a novel exploration pipeline where a VLM performs high-level strategic decision-making, guiding a conventional low-level robotics control stack. At decision points, the robot generates a multimodal prompt with its current map and visual imagery of potential paths, or frontiers. The VLM analyzes this prompt to select the most promising frontier, replacing simple geometric heuristics with contextual spatial reasoning. This approach, validated in simulation across six indoor environments, improves map coverage by up to 24\% over existing methods. Our pipeline is lightweight, training-free, and easily transferable to any robot with standard sensors and an internet connection.

The deployment of large language models (LLMs) on resource-constrained devices remains challenging, spurring interest in split inference, where models are partitioned between client and server to reduce computational burden and enhance privacy by transmitting only intermediate activations. However, the privacy-preserving capabilities of split inference, particularly in the context of LLMs, have not been exhaustively investigated. To fill this gap, we introduce ActInv, which solves an intermediate activation matching problem to reconstruct the client's input. Extensive evaluations demonstrate that ActInv achieves high-fidelity reconstructions, even in the presence of common perturbation-based defenses such as Gaussian noise injection and activation sparsification. To systematically understand this vulnerability, we develop Perturbation Amplification Factor (PAF), a metric for quantifying a layer's inherent resistance to reconstruction. Our analysis reveals that privacy vulnerability is not uniform across layers, with some layers being highly susceptible to leakage while others offer natural resistance. Furthermore, we demonstrate that defense effectiveness can be significantly improved by calibrating perturbation directions to maximize reconstruction error during backpropagation. Building on these insights, we design PriPert and conduct comprehensive evaluations, covering privacy, utility, and computational overhead, to demonstrate its effectiveness.

The attack surface of a multimodal large language model (MLLM) is language-dependent in ways that reveal the mechanistic structure of alignment failures. We present the first systematic cross-lingual, multimodal red-teaming study comparing jailbreak vulnerability in US English (en-US) and Mexican Spanish (es-MX) across four frontier MLLMs: Claude Sonnet 4.5, GPT-5, Pixtral Large, and Qwen Omni. Using a fixed adversarial benchmark of 363 diverse prompt scenarios administered in text-only and multimodal conditions, we collected 52,272 harm ratings and binary attack success judgements from matched panels of nine native-speaker annotators per language group. Our central finding is that language does not scale vulnerability uniformly. Bayesian mixed-effects analyses reveal that linguistic framing attacks such as role-play become substantially less effective under Spanish prompting, while visually explicit multimodal attacks become more effective, which directly implicates the prompt-language interface rather than global annotator leniency. This dissociation indicates that linguistic and visual alignment failures operate through distinct mechanisms, and that switching language is sufficient to expose that separation. The practical consequence is that safety rankings are not preserved across languages. Qwen Omni overtakes Pixtral Large as the most vulnerable model among es-MX participants, a rank reversal no scalar correction of English-condition scores could recover, and absolute attack success rates have declined across model generations without closing the gaps between them. These findings demonstrate that safety evaluation frameworks treating language and modality as independent dimensions fundamentally misspecify the attack surface of globally deployed MLLMs, and must be redesigned accordingly.

Role prompts of the form As X, do Y admit a clean linear decomposition at one specific site in the residual stream: the prompt-to-answer transition -- the last prompt token together with the first two generated tokens -- in an early/mid layer band. There, persona and task contribute through partially orthogonal additive directions. Forming a pure persona effect $Δ_X$, a pure task effect $Δ_Y$, and substituting $h_{BB} + Δ_X + Δ_Y$ for the clean residual yields downstream output within a small KL of clean on Gemma-2-2B-IT and Qwen-2.5-\{1.5B, 3B\}-Instruct, across a 12-cell short grid and a 48-cell long-persona grid, with persona-specific behavioral markers preserved. The natural inference from this additive structure is that the role prompt can be compressed into a single cached residual vector. \emph{We show it cannot.} Injecting the cached additive prediction -- or even the oracle clean residual $h_{XY}$ -- into a baseline host prompt with the persona text removed does not approach the clean long-persona target, at one site or at many layers. Persona-conditioned multi-token generation flows through attention back to the persona-text positions throughout the prompt, which no residual at one site reproduces. Local additivity in the residual stream does not imply prompt compressibility. The additive structure at the prompt-to-answer transition supports interpretability and fine-grained steering of persona or task contributions; persona-conditioned behavior across the full continuation depends on a distributed prompt/KV mechanism that local activation arithmetic does not displace.

I present Lepton (Letter Prediction), a fine-tuned BERT classifier that predicts whether a title in a Classical Chinese wenji table of contents is a personal letter or a closely confusable preface (particularly the farewell-preface). Lepton fine-tunes bert-base-chinese on 5438 hand-labeled wenji titles from thirty-three late-Ming and early-Qing literati. I've deployed the model on Hugging Face and has been used at the China Biographical Database (CBDB) to identify approximately fifty-five thousand letters across mid-Ming through early-Qing wenji, populating the Ming Letter Platform.

Topic modeling in applied psychology increasingly spans two methodological traditions: probabilistic bag-of-words models and newer embedding-based approaches. Yet many evaluations of these methods rely on longer and cleaner benchmark corpora, leaving less guidance for short, open-ended survey responses. This paper compares Structural Topic Models (STM), a probabilistic topic model, and BERTopic, an embedding-based model, for analyzing open-ended survey responses. We evaluated three STM conditions and five BERTopic conditions, varying typographical correction, stemming, embedding choice, and contextual augmentation, a strategy we introduced to provide additional semantic context for very short responses. Results indicate that BERTopic consistently produced higher topic coherence than STM, with contextual augmentation yielding the strongest performance gains. In contrast, higher-dimensional embeddings alone did not improve coherence and were associated with greater data loss. Qualitative evaluation showed that BERTopic generated more interpretable and stable topics, while STM topics were often broader and more mixed. However, STM provides stronger support for inferential covariate analysis, whereas BERTopic covariate comparisons are primarily descriptive. These findings suggest that STM and BERTopic offer complementary strengths. We conclude with practical guidance for selecting and combining topic modeling approaches in applied social science research.

Mixture-of-Experts Large Language Models (MoE-LLMs) achieve strong performance but incur substantial memory overhead due to massive expert parameters. Mixed-precision quantization mitigates this cost by allocating expert-wise bit-widths based on their importance, approaching the accuracy-memory Pareto frontier and enabling extreme low-bit quantization. However, existing methods rely on layer-wise importance estimation and overlook router shifts induced by quantization, resulting in suboptimal allocation and routing. In this work, we propose Global Expert-level Mixed-precision Quantization (GEMQ) to overcome these limitations via (1) a global linear-programming formulation that captures model-wide expert importance based on quantization error analysis, and (2) efficient router fine-tuning to adapt routing to quantized experts. These components are integrated into a progressive quantization framework that iteratively refines importance estimation and allocation. Experiments demonstrate that GEMQ significantly reduces memory and accelerates inference with minimal accuracy degradation. Source code is available at https://github.com/jndeng/GEMQ .

Large language models (LLMs) increasingly rely on long-context processing, but expanding context windows introduces substantial computational and financial costs. Existing context reduction approaches, including retrieval and memory compression methods, are typically evaluated using performance and efficiency metrics independently, limiting systematic comparison and deployment-aware decision-making. This paper introduces The Efficiency Frontier, a unified framework for cost-performance optimization in LLM context management. The framework models context strategy selection as a deployment-aware optimization problem that jointly accounts for task performance, token cost, and preprocessing reuse through amortized cost modeling. Unlike existing evaluations that compare methods in isolation, the proposed framework enables decision-oriented analysis of when different context management strategies become preferable under varying operational conditions. Evaluated on 5,000 HotpotQA instances, the framework reveals distinct operational regimes and transition boundaries between retrieval-based and preprocessing-based strategies. Results show that deployment-aware optimization reduces effective token usage by approximately 25% at comparable performance ($F1 \approx 0.78$), while amortized memory compression achieves over 50% lower token cost relative to full-context prompting in higher-performance settings. Overall, the proposed framework provides a principled and practical foundation for evaluating and deploying scalable, efficient, and sustainable LLM systems.

Large language models (LLMs) are increasingly used across diverse linguistic and cultural contexts, yet their cultural knowledge remains uneven across regions and languages. We present the DFKI-MLT system for SemEval-2026 Task 7 on cultural awareness, where we apply activation steering to multilingual LLMs using language vectors extracted from parallel FLORES data. Our method performs inference-time adaptation by adding language-specific steering vectors to the residual stream at a selected transformer layer, without any parameter updates. We participated in both the short-answer (SAQ) and multiple-choice (MCQ) tracks; however, only our MCQ submission received an official score. In the official MCQ track, we achieved 86.96% accuracy, ranking 7th out of 17 teams. To better understand system behavior, we conduct post-hoc analyses on the shared-task MCQ and SAQ settings. These analyses show that activation steering yields modest and heterogeneous improvements on cultural reasoning: gains are strongly layer-sensitive, vary substantially across language-region pairs, with some configurations even degrading performance, and interact with prompt formulation, comparing generic and culturally conditioned prompts. Our findings suggest that prompt design and activation steering should be jointly optimized for culturally aware multilingual inference.

Reinforcement learning (RL) has emerged as a viable recipe for training LLM agents to reason over external memory banks in multi-session dialogue. Existing work trains exclusively on a single benchmark, leaving open how the composition of training data shapes the skills a memory agent acquires. We present a controlled empirical study that holds architecture, RL algorithm, and all hyperparameters fixed and varies only the training curriculum across three conditions: in-domain (LoCoMo), mixed-benchmark (LoCoMo + LongMemEval), and out-of-domain (LongMemEval only). Across two benchmarks and ten question types, curriculum composition acts as a fine-grained lever on specialization rather than a uniform scaling factor on performance. The mixed curriculum yields the strongest overall F1 on both evaluation sets. Training on a narrow out-of-domain set transfers a targeted skill - temporal reasoning - despite weak aggregate performance. Per-type differences substantially exceed aggregate differences, indicating that single-number benchmark comparisons systematically underreport curriculum effects. We further report two practical lessons from adapting GRPO to a single-GPU regime: cross-benchmark mixing requires filtering format-specific noise from memory banks to preserve training signal, and binary exact-match reward produces no learning signal at the small group sizes (G = 4) required on one GPU, motivating continuous reward functions in this regime.

ModeSwitch-LLM is a lightweight request-boundary controller for improving single-GPU large language model inference efficiency by routing each request to an appropriate fixed inference mode. Instead of relying on one static serving configuration, the system selects among FP16, quantized modes, speculative decoding, and hybrid modes such as GPTQ plus prefix caching and INT8 plus continuous batching using cheap workload-level features. We evaluate ModeSwitch-LLM on Meta-Llama-3.1-8B-Instruct served on a single NVIDIA A100 GPU. On deployment-style synthetic workloads, the online controller achieves a 2.10x mean latency speedup over FP16 and a 0.48x mean energy ratio, corresponding to 51.7% lower energy per token. On automatic benchmarks used as a quality gate, accuracy remains close to FP16 with a mean delta of +0.17 percentage points. We also evaluate lightweight learned routers, but find that they do not clearly outperform the rule-based controller because they add routing overhead and more often select modes that violate quality, energy, or memory constraints. These results show that simple request-aware routing can recover substantial efficiency from existing inference modes without retraining the model or changing its architecture.

Model collapse, the progressive degradation of LLMs trained on their own outputs, has been characterized statistically but lacks a linguistic explanation for which structures degrade, in what order, and why. We show that iterated learning theory from cultural evolution fills this gap. We derive five falsifiable predictions, distinguish those uniquely discriminative for the theory from confirmatory ones, and test them by self-training LLaMA-2-7B and Mistral-7B over 10 generations in English, German, and Turkish. The critical discriminative finding: compositionality follows a non-monotonic trajectory (initially rising, then falling) under unfiltered self-training. This signature persists with maximally regular seed data (ruling out noise removal) and is sustained only by task-grounded filtering, not random filtering, providing the first LLM-scale evidence for the compression-communication tradeoff. All predictions are confirmed with large effect sizes (Hedges' $g > 1.6$; $\mathrm{BF}_{10} > 100$), and LLM regularization gradients closely match human behavioral data ($R^2 = 0.94$). These results reframe model collapse as a cultural transmission phenomenon and yield concrete principles for self-training pipeline design.

We present DreamerNLplus, a hybrid framework for modeling mental health dynamics from social media timelines in the CLPsych 2026 shared task. Our system addresses three tasks: psychological state modeling, temporal change detection, and sequence-level summarization. For Task 1, we combine LLM-based data augmentation, DeBERTa classification, and Random Forest regression for structured state prediction. For Task 2, we use few-shot prompting with a locally deployed Llama 3.1 model to detect Switch and Escalation events using short-term temporal context. For Task 3.1, we explore both a deterministic rule-based summarization pipeline and a few-shot LLM-based approach, ranking \textbf{2nd} officially. Our RAG-based method achieves strong performance in Task 3.2, ranking \textbf{1st} for Improvement and \textbf{3rd} for Deterioration, demonstrating its ability to capture recurrent psychological change patterns across timelines. Our analysis reveals key challenges, including the mismatch between classification and regression performance, the difficulty of modeling temporal transitions, and the disagreement between semantic and similarity-based evaluation metrics. These findings highlight the complexity of modeling mental health dynamics and motivate future work on unified evaluation frameworks. We share our code and prompts at https://github.com/4dpicture/CLPsych2026

Agentic text-simulation systems write in sequence, with each item becoming possible context for later steps. That makes uncertainty path-dependent: an early ambiguity can affect later outputs. This paper studies this problem with HawkesLLM, a framework that separates temporal influence modeling from text generation. We represent the cascade as a network whose nodes are text-generating agents. A multivariate Hawkes process models how these nodes activate over time and which earlier node outputs should influence later prompts. A language model then writes each new event from the compact memory selected by this temporal model. We evaluate the framework on a held-out Global Database of Events, Language, and Tone (GDELT) news-cascade case study. The diagnostics track semantic alignment with local held-out references and separate local drift from global drift. In this setting, HawkesLLM improves late-stage semantic alignment under a compact prompt-memory budget.

How do learners acquire knowledge of what is unacceptable without negative evidence? Construction Grammar proposes statistical preemption: exposure to a conventional form (e.g., "donated the books to the library") preempts structurally possible but unattested alternatives ("*donated the library the books"). We present a computational study that, for the first time, directly dissociates statistical preemption from the competing entrenchment hypothesis in large language models within a single converging design. Across four experiments spanning 120 English verb-construction pairings (dative, causative, locative), we show that (1) LLM surprisal patterns correlate strongly with human acceptability judgments ($r = 0.79$), validated against three independent behavioral datasets; (2) these patterns are driven by competing-form frequency rather than overall verb frequency, confirmed by non-circular partial correlations; (3) preemption sensitivity scales as a power law with model size; and (4) a controlled fine-tuning intervention causally demonstrates that manipulating competing-form frequencies shifts preemption behavior in the predicted direction, with reverse-direction controls ruling out frequency-sensitivity confounds. These results provide converging evidence that neural language models acquire negative linguistic knowledge through distributional competition, the core mechanism posited by Construction Grammar.

Sparse autoencoders (SAEs) enable feature-level mechanistic interpretability and activation steering in large language models (LLMs), but SAE-based language control remains unreliable in multilingual settings: most SAEs are trained on English-only data, and steering layers are chosen heuristically. We address these limitations by advancing a principled, mechanistic account of multilingual language steering with SAEs. First, we show that training SAEs on multilingual data consistently strengthens cross-lingual representations and yields more reliable, quality-preserving language control across layers and model families. Second, we introduce an \emph{a priori} steering layer-selection rule based on the intersection of multilingual alignment and language separability, which predicts effective intervention depths without exhaustive layerwise search. We evaluate our approach on LLaMA-3.1-8B and Gemma-2-9B across machine translation and cross-lingual summarization (CrossSumm), using SpBLEU, ROUGE-L, COMET, and LaSE. Our results show that multilingual SAEs combined with intersection-selected layers stabilize the trade-off between language identification accuracy and generation quality, providing a principled, predictive, representation-level account of multilingual SAE steering.

Intermediate layers of large language models (LLMs) best predict human brain responses to language, one of the most robust findings in computational neurolinguistics, yet why remains mechanistically unexplained. We address this gap by bridging sparse autoencoders (SAEs) from mechanistic interpretability with neural encoding models, decomposing GPT-2 XL and Llama-3.1-8B into 16K-32K interpretable features per layer. A human-validated taxonomy ($κ\geq 0.74$) reveals that semantic features alone recover 94% of peak encoding performance ($r=0.285$), substantially exceeding variance-matched baselines ($p<0.001$, $d=1.31$). Beyond this aggregate dominance, we test a novel cortical topography prediction: five semantic subcategories derived a priori from three independent neuroscience programs should map onto distinct brain regions. A formal convergence test confirms this alignment (Spearman $ρ=0.72$, $p<0.001$; hypergeometric $p=0.007$), demonstrating that SAE-discovered features recapitulate known cortical semantic organization at a granularity inaccessible to prior methods. SAE features further predict human reading times beyond lexical controls ($Δ\mathrm{logLik}=38.4$, $p<0.001$), and an exploratory prediction-error analysis provides preliminary evidence that the brain additionally encodes unexpected semantic content. Results generalize across English, Chinese, and French.

Brain-LLM alignment is well established in English, yet the brain's language network is neuroanatomically universal across languages. Does alignment also generalize cross-linguistically, and what governs the variation? We test this using fMRI data from 112 participants across English, Chinese, and French (the Le Petit Prince corpus) and seven LLMs spanning English-dominant, Chinese-dominant, and multilingual architectures. Our central finding is that training-language dominance, not an inherent property of English, drives the alignment pattern: a Chinese-dominant model (Baichuan2-7B), architecture-matched to LLaMA-2-7B, reverses the gradient entirely, aligning best with Chinese brains and worst with English. Beyond training dominance, formal typological distance independently covaries with alignment degradation, syntax-associated brain regions (IFG) show $2.3\times$ steeper typological gradients than lexico-semantic regions (PTL), and tokenization fertility accounts for $\sim$60% of a cross-linguistic shift in optimal encoding layer. These results reveal that the apparent "English advantage" in brain-LLM alignment is an artifact of training data composition, while the remaining variation reflects genuine typological structure concentrated in syntactic processing.

Machine learning methods rely on data. However, gathering suitable data can be challenging due to availability constraints, cost, or the need for domain expertise. Expanding datasets with additional sources is a common response to limited data, yet this practice does not always improve downstream performance and can sometimes lead to a loss of performance, known as negative transfer. We propose RADAR, a simple, geometrically grounded metric for estimating cross-domain transferability in foundation models. RADAR analyzes the layer-wise evolution of representations by measuring angular alignments and relative changes in distance along layer-to-layer displacement trajectories, and by comparing empirical distributions of within-domain and cross-domain dynamics. We hypothesize that domain transferability is related to the divergence between these trajectory distributions. We evaluate the metric across multiple modalities, including cross-lingual sentiment classification with text embedding models and cross-domain image classification with foundation vision models. Across several settings, RADAR provides competitive predictive performance relative to existing transferability metrics on several vision and text benchmarks, with particularly strong results when domain transitions are smooth or cleanly separated. Our ablations further suggest that the effectiveness of transferability estimation depends on the geometry of the model's internal representation space, with different modalities favoring different topological formulations.

Large language models now write software, draft legal documents, and produce clinical notes, yet fundamental limits, from Turing and Arrow to the No Free Lunch theorems, shape what computation can do. This thesis turns such impossibility results from curiosities into design rules. Its flagship result proves an accuracy ceiling set by architecture alone: past a critical reasoning depth, no amount of training moves it, at any adapter rank, sample size, or loss function. Computable before deployment from layer count and embedding width, this Deterministic Horizon is measured between nineteen and thirty-one across twelve transformer architectures, and fine-tuning on optimal-length traces recovers under four percentage points. The mechanism is a capacity invariant of the residual stream, and an information-theoretic conversion yields super-exponential accuracy decay past the horizon. An unconditional circuit-complexity lower bound for modular exponentiation against constant-depth prime-modulus circuits complements this result. The same argument recasts across subfields: preference learning under any misspecified model jumps discontinuously in sample complexity; multi-stage retrieval pipelines require at least as many independent metrics as stages; standard truthful auctions fail for agents with prompt-dependent valuations; and zero-knowledge verification of neural inference pays a measured overhead of one hundred ten to one hundred ninety times per non-linear activation. Together these form a catalogue of sixteen specifications, each pairing a computable boundary, a quantified violation cost, and a constructive design rule: two compositions are proved, one pairing is an honest obstruction, and four remain open. The impossibility-specification methodology is offered for the generative research programme that trustworthy AI may need. Every fundamental limit of AI is also a design rule.

Characteristic linguistic behaviors associated with Social Language Disorder (SLD) in autism spectrum disorder, including echoic repetition, pronoun displacement, and stereotyped media quoting, are largely absent from spontaneous conversation and only emerge under specific conversational conditions. In structured clinical assessments, this latency means that questioning strategy selection is a critical yet underappreciated determinant of how much diagnostic information a conversation yields. Whether large language models (LLMs) can be guided to proactively select questioning strategies that systematically surface these latent traits remains largely unexplored. Here we present TPA (Think, Plan, Ask), a proactive multi-agent dialogue framework applied to the language assessment component of the Autism Diagnostic Observation Schedule Module 4 (ADOS-2), in which a doctor agent explicitly reasons about which traits remain unobserved before selecting a clinically grounded strategy and generating a targeted question. A patient agent grounded in real ADOS-2 clinical data enables reproducible evaluation without real patient participation, validated across three independent experiments confirming adequate fidelity to real patient language. Evaluated on 484 episodes from 35 patients, TPA outperforms six competitive dialogue planning baselines across all primary metrics, achieving 82.1% SLD trait coverage, 16.6% higher than automated replay of real clinical dialogues conducted by trained clinicians (65.5%), with substantially greater per-turn diagnostic efficiency (AUCC: 0.628 vs. 0.458, absolute gain +0.170). These results demonstrate that proactive questioning strategy selection substantially improves the efficiency of automated SLD trait assessment, with direct implications for scalable AI-assisted clinical screening.

Fill-in-the-middle (FIM) is a pretraining objective widely used to equip causal language models with infilling ability, yet its effect on verbatim memorization remains underexplored. We study the memorization dynamics of FIM in a controlled setting by pretraining matched Llama 3.2 models with FIM and standard left-to-right (LTR) objectives on a FineWeb-Gutenberg corpus containing repeated Gutenberg excerpts. With prefix-based probes, FIM more often recovers short or partially matching spans, while LTR more often assigns high confidence to long exact continuations. We observe that verbatim extraction under FIM-training grows approximately linearly with repetitions over the tested range. Evaluating native FIM-format probes reveals that suffix context is not sufficient: verbatim recall under FIM-training remains strongly anchored in prefix context. Our results also show that evaluating only one span length or probing format can miss important nuances in memorization behavior.

Employees often struggle to identify ``who knows what,'' leading to organizational productivity losses. We investigate whether Large Language Models (LLMs) can infer individual domain knowledge directly from long-term Slack logs. Analyzing 27,188 messages from 43 users, we evaluated seven models (including Gemini, Claude, and GPT families) by comparing their zero-shot estimates against self-reported skill ratings from 27 participants. Gemini 2.5 Flash achieved the lowest error (MAE 21.13%), while GPT models showed significantly larger discrepancies. Notably, estimation accuracy depended only weakly on message volume, indicating that more text alone does not guarantee better inference. These findings demonstrate the feasibility and current limits of automated expertise mapping, highlighting the need for privacy-preserving deployments and richer, structure-aware representations of human knowledge.

Large Language Models (LLMs) are optimized to produce distributionally plausible continuations rather than to explicitly verify whether generated propositions are entailed by source documents. This inductive bias enables generalization, but it does not encode whether responses are grounded with respect to a reference. These issues limit the use of LLMs in domains where strict factual correctness is crucial, such as clinical decision support. Existing hallucination detection approaches improve factuality through retrieval augmentation, self-consistency, or claim verification, but generally do not learn directly over alignment topology. To leverage alignment topology as an inductive bias, we construct aligned bipartite graphs between reference information and LLM outputs and train a graph neural network (GNN) to model alignment structure using message passing. The method achieves state-of-the-art results on four diverse hallucination and question-answering datasets, outperforming all compared methods, including foundational LLMs such as GPT-4o.

While multi-task learning based automatic radiology report generation (RRG) is widely adopted to ensure clinical consistency, most focus on architectural designs yet remain limited to coarse linear scalarization strategies. These strategies cannot effectively balance the hard constraints of discriminative clinical supervision with the smoothness requirements of report generation. To address these problems, we analyze the failure mechanism of linear scalarization from the perspective of gradient dynamics, utilizing the stochastic differential equation (SDE) framework to characterize it as a "Double Dilemma" of drift term deviation and diffusion term decay. Based on this, we propose a backbone-agnostic optimizer named Conflict-Averse Magnitude-Enhanced Gradient Descent (CAME-Grad). Through conflict-averse direction rectification and magnitude-enhanced energy injection, the algorithm not only ensures geometric validity, but also avoids local optimal solutions. Then, the adaptive gradient fusion mechanism is used to establish a dynamic balance between the theoretical optimal direction and the task-specific inductive bias. Experiments show that as a universal plug-and-play optimizer, CAME-Grad brings substantial and consistent improvements across eight diverse RRG methods, elevating overall clinical efficacy performance by an average of 2.3% on MIMIC-CXR and 1.9% on IU X-Ray. Our code is available at https://github.com/vpsg-research/CAME-Grad.

Safety classifiers are essential safeguards within generative AI systems, filtering harmful content or identifying at-risk users when interacting with large language models. Despite their necessity, these models are trained on sensitive datasets including discussions of self-harm and mental health, raising important, yet poorly understood, privacy concerns. Membership inference attacks (MIAs) allow adversaries to infer membership of examples used to train models. In this work, we hypothesize that identifying the examples on which the classifier is least confident are informative for an adversary to infer membership. This reflects a localized failure of generalization, where the model relies on memorization to resolve ambiguity in the training set. To investigate this, we introduce a new boundary-targeted selection strategy that identifies low confidence examples that amplify the signal of an examples membership within a training set. Our experimental results show that an adversary can recover 19% of the conversations a safety classifier flagged as indicating user distress, at a 5% false-positive rate, on a classifier fine-tuned for detecting a user who may require emotional support. This is $3.5$ times more than attacking using state-of-the-art MIA methods alone. Finally, we characterize the boundary laying examples and show that content-based filtering is ineffective for protection, and existing noise strategies can effectively mitigate susceptibility of these examples.

The rapid progress of large language models (LLMs) is shifting semantic search toward a question-answering paradigm, where users ask questions and LLMs generate responses. In high-stake domains such as law, retrieval-augmented generation (RAG) is commonly used to mitigate hallucinations in generated responses. Nonetheless, prior work shows that RAG systems, whether general-purpose or legal-specific, still hallucinate at varying rates, making fine-grained evaluation essential. Despite the need, existing evaluation frameworks for legal RAG systems lack the granularity required to provide detailed analysis of retrieval and generation performance separately. Moreover, current benchmarks are largely English-only and centered on legal expert queries, overlooking non-expert needs. We introduce ClaimRAG-LAW, a comprehensive dataset for legal RAG that supports French and English, targets both experts and non-experts, and includes diverse question types reflecting realistic scenarios. We further apply a fine-grained evaluation framework of state-of-the-art legal RAG systems, revealing limitations in retrieval, generation, and claim-level analysis in the legal domain.

Neural morphological generation systems often achieve high aggregate accuracy on benchmark datasets, yet such performance can conceal systematic errors concentrated in rare morphological subclasses. We examine Japanese past-tense verb inflection and show that a very small, structurally specific irregular subtype (<1% of data) accounts for a disproportionate share of model errors. Controlled ablation experiments demonstrate that removing this subtype yields larger improvements in generalization than removing all irregular verbs, indicating that not all irregularity contributes equally to model instability. These findings suggest that error concentration is driven by the interaction between extreme low-frequency morphological patterns and specific morphophonological processes, particularly gemination. We argue that morphological evaluation should incorporate finer-grained subclass analysis beyond standard conjugation categories.

Recent large language models support inputs of up to 10 million tokens, yet they perform poorly on long-context tasks that require complex reasoning. Such tasks can be solved using only a subset of the input -- a proxy context -- rather than the full sequence. Despite sharing the same underlying reasoning process, models exhibit a significant performance disparity between proxy and full contexts. To improve long-context reasoning, we propose ProxyCoT, a novel training framework that transfers reasoning capabilities from short proxy contexts to full long contexts. Specifically, we first obtain high-quality chain-of-thought reasoning traces on proxy contexts through reinforcement learning or distillation from a larger teacher model, and then ground the generated traces in full long contexts with supervised fine-tuning. Experiments across different datasets demonstrate that ProxyCoT consistently outperforms strong baselines with reduced computational overhead. Furthermore, models trained with ProxyCoT generalize their long-context reasoning capabilities to out-of-domain tasks.

Decentraland, a decentralized virtual reality platform operating within the expanding Metaverse ecosystem, utilizes its native MANA token to facilitate virtual asset transactions and governance. This study investigates the integration of Discord community sentiment with multi-modal financial data to enhance cryptocurrency price prediction within virtual world economies. We address: (1) identifying sentiment patterns within Decentraland's Discord community, and (2) evaluating the impact of multi-modal features on token return forecasting. Using a BERT-based large language model for sentiment analysis, we develop two LSTM architectures: a baseline incorporating historical prices and a multi-modal variant integrating sentiment scores, trading volume, and market capitalization. Results indicate predominantly neutral community sentiment with a positive skew. The multi-modal model significantly outperforms the price-only baseline in prediction accuracy. These findings demonstrate the predictive value of community-derived signals for virtual economy forecasting and establish a foundation for future research at the intersection of immersive virtual environments, natural language processing, and cryptocurrency market analysis.

Conversational AI has now reached billions of users, yet existing datasets capture only what people say, not what they think. We introduce ThoughtTrace, the first large-scale dataset that pairs real-world multi-turn human--AI conversations with users' self-reported thoughts: their reasons for sending prompts and reactions to assistant responses. ThoughtTrace comprises 1,058 users, 2,155 conversations, 17,058 turns, and 10,174 thought annotations collected across 20 language models. Our analysis shows that ThoughtTrace captures long-horizon, topically diverse interactions, and that thoughts are semantically distinct from messages, difficult for frontier LLMs to infer from context, diverse in content, and tied to conversation stages. We further demonstrate the utility of thoughts for downstream modeling. First, thoughts improve user-behavior prediction as inference-time context. Second, thought-guided rewrites provide fine-grained alignment signals for training personalized assistants. Together, ThoughtTrace establishes user thoughts as a new data modality for studying the cognitive dynamics behind human--AI interaction and provides a foundation for building assistants that better understand and adapt to users' latent goals, preferences, and needs.

We present an orthography-aware error analysis of Japanese past-tense morphological inflection, treating hiragana not merely as a transcriptional medium, but as a representational system encoding morphophonological distinctions that may influence model generalization. We evaluate two character-level sequence-to-sequence architectures on past-tense formation using datasets formatted according to the SIGMORPHON 2020 and 2023 shared task conventions. Despite high aggregate accuracy, models exhibit systematic, linguistically interpretable errors that cluster around specific orthographic properties of hiragana. We introduce a concise error taxonomy capturing seven primary failure modes and provide both quantitative and qualitative analyses. Gemination-related errors dominate residual failures, accounting for 75-80% of errors, particularly in verbs whose stems end in the vowel e and require gemination before the past-tense suffix. Error patterns remain highly consistent across architectures and random seeds, suggesting a robust interaction between orthographic representation, morphological structure, and data frequency effects in shaping model generalization. These results underscore the necessity of orthography-aware evaluation for understanding neural generalization in morphologically complex languages.

Large language models (LLMs) are increasingly being applied to financial analysis, reporting, investment decision support, risk management, compliance, and professional training. However, robust evaluation of their domain competence in finance remains incomplete. Widely used open benchmarks such as FinQA, ConvFinQA, and TAT-QA have played an important role in advancing financial question answering and numerical reasoning, but they focus primarily on question answering over financial reports and do not provide an explicit hierarchy of professional difficulty. Broader resources, including FinanceBench, PIXIU, FinBen, and FLaME, expand the coverage of financial tasks, yet the problem of evaluating the transition from foundational knowledge to expert-level financial reasoning remains open. In this work, we present FINESSE-Bench, a suite of eight specialized benchmarks comprising 3,993 questions for hierarchical evaluation of financial competencies in LLMs. FINESSE-Bench combines exam-oriented datasets inspired by professional certifications (CFA-like Levels 1-3, CMT-like Level 2, and CFTe-like Level 1), applied trading task collections, and a Russian-language olympiad benchmark. This design enables evaluation of domain breadth, performance degradation as difficulty increases, the ability to solve computational tasks, and model behavior in specialized financial domains. We also describe a unified evaluation protocol covering multiple-choice questions, numerical answers, and short open-ended responses, together with an automated scoring scheme for freeform answers based on the LLM-as-judge paradigm. FINESSE-Bench is intended both as a complement to existing open financial benchmarks and as a tool for more substantive evaluation of professionally relevant financial competencies in large language models.

Recent advances in vision-language models have enabled mobile GUI agents to perceive visual interfaces and execute user instructions, but reliable prediction of action consequences remains critical for long-horizon and high-risk interactions. Existing mobile world models provide either text-based or image-based future states, yet it remains unclear which representation is useful, whether generated rollouts can replace real environments, and how test-time guidance helps agents of different strengths. To answer the above questions, we filter and annotate mobile world-model data, then train world models across four modalities: delta text, full text, diffusion-based images, and renderable code. These models achieve SoTA performance on both MobileWorldBench and Code2WorldBench. Furthermore, by evaluating their downstream utility on AITZ, AndroidControl, and AndroidWorld, we obtain three findings. First, renderable code reconstruction achieves high in-distribution fidelity and provides effective multimodal supervision for data construction, while text-based feedback is more robust for online out-of-distribution (OOD) execution. Second, world-model-generated trajectories can provide transferable interaction experience in the training process and improve agents' end-to-end task performance, although these data do not preserve the original distribution. Last, for overconfident mobile agents with low action entropy, posterior self-reflection provides limited gains, suggesting that world models are more effective as prior perception or training supervision than as universal post-hoc verifiers.

Large Language Models (LLMs) are increasingly used as proxies for human perception in urban analysis, yet it remains unclear whether persona prompting produces meaningful and reproducible behavioral diversity. We investigate whether distinct personas influence urban sentiment judgments generated by multimodal LLMs. Using a factorial set of personas spanning gender, economic status, political orientation, and personality, we instantiate multiple agents per persona to evaluate urban scene images from the PerceptSent dataset and assess both within-persona consistency and cross-persona variation. Results show strong convergence among agents sharing a persona, indicating stable and reproducible behavior. However, cross-persona differentiation is limited: economic status and personality induce statistically detectable but practically modest variation, while gender shows no measurable effect and political orientation only negligible impact. Agents also exhibit an extremity bias, collapsing intermediate sentiment categories common in human annotations. As a result, performance remains strong on coarse-grained polarity tasks but degrades as sentiment resolution increases, suggesting that simple label-based persona prompting does not capture fine-grained perceptual judgments. To isolate the contribution of persona conditioning, we additionally evaluate the same model without personas. Surprisingly, the no-persona model sometimes matches or exceeds persona-conditioned agreement with human labels across all task variants, suggesting that simple label-based persona prompting may add limited annotation value in this setting.

Maintaining information in context is essential in successful real-time language comprehension, but maintenance is cognitively costly and can slow processing. We hypothesize that rational language users selectively maintain information that is crucial for future prediction, guided by syntactic structure. Under this view, two factors affect maintenance cost: the number of predicted heads and the number of incomplete dependencies. Although these factors have been treated as competing hypotheses in the literature, our account predicts that they are not reducible to one another. We show this is the case in a naturalistic reading time dataset in Japanese, a language in which the two factors contrast particularly clearly. We further show that there is a tradeoff such that readers that slow down for maintenance tend to benefit more from predictability, providing additional support for the proposed account. These patterns are not evident in English, however, and we highlight some issues to be resolved to understand the contribution of syntax in memory-efficient processing of various languages.

Real-time detection and mitigation of technical anomalies are critical for large-scale cloud-native services, where even minutes of downtime can result in massive financial losses and diminished user trust. While customer incidents serve as a vital signal for discovering risks missed by monitoring, extracting actionable intelligence from this data remains challenging due to extreme noise, high throughput, and semantic complexity of diverse business lines. In this paper, we present TingIS, an end-to-end system designed for enterprise-grade incident discovery. At the core of TingIS is a multi-stage event linking engine that synergizes efficient indexing techniques with Large Language Models (LLMs) to make informed decisions on event merging, enabling the stable extraction of actionable incidents from just a handful of diverse user descriptions. This engine is complemented by a cascaded routing mechanism for precise business attribution and a multi-dimensional noise reduction pipeline that integrates domain knowledge, statistical patterns, and behavioral filtering. Deployed in a production environment handling a peak throughput of over 2,000 messages per minute and 300,000 messages per day, TingIS achieves a P90 alert latency of 3.5 minutes and a 95\% discovery rate for high-priority incidents. Benchmarks constructed from real-world data demonstrate that TingIS significantly outperforms baseline methods in routing accuracy, clustering quality, and Signal-to-Noise Ratio.

Reinforcement learning with verifiable rewards (RLVR) has significantly advanced the reasoning ability of vision-language models (VLMs). However, the inherent text-dominated nature of VLMs often leads to insufficient visual faithfulness, characterized by sparse attention activation to visual tokens. More importantly, our empirical analysis reveals that temporal visual forgetting along reasoning steps exacerbates this deficiency. To bridge this gap, we propose Visually-Guided Policy Optimization (VGPO), a novel framework to reinforce visual focus during policy optimization. Specifically, VGPO initially introduces a Visual Attention Compensation mechanism that leverages visual similarity to localize and amplify visual cues, while progressively elevating visual expectations in later steps to counteract visual forgetting. Building on this mechanism, we implement a dual-grained advantage re-weighting strategy: the intra-trajectory level highlights tokens exhibiting relatively high visual activation, while the inter-trajectory level prioritizes trajectories demonstrating superior visual accumulation. Extensive experiments demonstrate that VGPO achieves better visual activation and superior performance in mathematical multimodal reasoning and visual-dependent tasks. The code has been released at https://github.com/wzb-bupt/VGPO.

Extracting structured information from academic PDF documents is non trivial: a single page typically combines free text metadata with tabular regions, exhibits cross program variation, and is susceptible to Unicode encoding artifacts that interfere with downstream parsing. This study evaluates the reliability of information extraction approaches for tabular PDF documents, using academic course registration documents (Kartu Rencana Studi or KRS) from Indonesian higher education as a case study. Three strategies are compared: LLM only, Hybrid Deterministic - LLM (regex & LLM), and a Camelot based pipeline with LLM fallback. Experiments were conducted on 140 documents for the LLM based test and 860 documents for the Camelot based pipeline evaluation, covering four study programs with varying data in tables and metadata. Three 12 - 14B LLM models (Gemma 3, Phi 4, and Qwen 2.5) were run locally using Ollama and a consumer grade CPU without a GPU. Evaluations used exact match (EM) and Levenshtein similarity (LS) metrics with a threshold of 0.7. Although not applicable to all models, the results show that the hybrid approach can improve efficiency compared to LLM only, especially for deterministic metadata. The Camelot based pipeline with LLM fallback produced the best combination of accuracy (EM and LS up to 0.99 - 1.00) and computational efficiency (less than 1 second per PDF in most cases). The Qwen 2.5:14b model demonstrated the most consistent performance across all scenarios. These findings confirm that integrating deterministic and LLM based methods is a reliable and efficient strategy for information extraction from tabular text based PDF documents in computationally constrained environments.

Transformer-based embedding models frequently exhibit geometric pathologies, such as anisotropy and length-induced representation collapse, which can degrade downstream retrieval performance. While prior work often attributes these issues directly to text length or attention mechanisms, we argue that the fundamental drivers are instead the inherent pooling operations coupled with internal semantic shift. In this paper, we establish a unified theoretical framework proving that contextual pooling intrinsically causes embedding collapse. Specifically, we mathematically prove that pooling semantically diverse sentences inevitably leads to micro-level semantic dilution, and strictly reduces the Mean Pairwise Distance of the vector space, guaranteeing macro-level spatial concentration. Grounded in these geometric insights, we formally define semantic shift to capture the natural semantic evolution and dispersion within a text. Through carefully controlled experiments across diverse models and corpora, we disentangle text length from semantic content. We demonstrate that semantic shift is the primary predictor of severe embedding concentration. Crucially, our retrieval evaluations reveal that anisotropy is fundamentally harmful only when induced by strong semantic shifts, reconciling conflicting observations in prior literature and offering a principled explanation for the long-context challenges faced by modern embedding models.

We evaluate Large Language Models (LLMs) in repeated game-theoretic settings to assess whether strategic performance reflects genuine reasoning or reliance on memorized patterns. We consider two canonical games, Prisoner's Dilemma (PD) and Rock-Paper-Scissors (RPS), upon which we introduce counterfactual variants that alter payoff structures and action labels, breaking familiar symmetries and dominance relations. Our multi-metric evaluation framework compares default and counterfactual instantiations, showcasing LLM limitations in incentive sensitivity, structural generalization and strategic reasoning within counterfactual environments.

Political speakers often avoid answering questions directly while maintaining the appearance of responsiveness. Despite its importance for public discourse, such strategic evasion remains underexplored in Natural Language Processing. We introduce SemEval-2026 Task 6, CLARITY, a shared task on political question evasion consisting of two subtasks: (i) clarity-level classification into Clear Reply, Ambivalent, and Clear Non-Reply, and (ii) evasion-level classification into nine fine-grained evasion strategies. The benchmark is constructed from U.S. presidential interviews and follows an expert-grounded taxonomy of response clarity and evasion. The task attracted 124 registered teams, who submitted 946 valid runs for clarity-level classification and 539 for evasion-level classification. Results show a substantial gap in difficulty between the two subtasks: the best system achieved 0.89 macro-F1 on clarity classification, surpassing the strongest baseline by a large margin, while the top evasion-level system reached 0.68 macro-F1, matching the best baseline. Overall, large language model prompting and hierarchical exploitation of the taxonomy emerged as the most effective strategies, with top systems consistently outperforming those that treated the two subtasks independently. CLARITY establishes political response evasion as a challenging benchmark for computational discourse analysis and highlights the difficulty of modeling strategic ambiguity in political language.

Natural language processing for the Turkic language family, spoken by over 200 million people across Eurasia, remains fragmented, with most languages lacking unified tooling and resources. We present TurkicNLP, an open-source Python library providing a single, consistent NLP pipeline for Turkic languages across four script families: Latin, Cyrillic, Perso-Arabic, and Old Turkic Runic. The library covers tokenization, morphological analysis, part-of-speech tagging, dependency parsing, named entity recognition, bidirectional script transliteration, cross-lingual sentence embeddings, and machine translation through one language-agnostic API. A modular multi-backend architecture integrates rule-based finite-state transducers and neural models transparently, with automatic script detection and routing between script variants. Outputs follow the CoNLL-U standard for full interoperability and extension. Code and documentation are hosted at https://github.com/turkic-nlp/turkicnlp .

We introduce BURMESE-SAN, the first holistic benchmark that systematically evaluates large language models (LLMs) for Burmese across three core NLP competencies: understanding (NLU), reasoning (NLR), and generation (NLG). BURMESE-SAN consolidates seven subtasks spanning these competencies, including Question Answering, Sentiment Analysis, Toxicity Detection, Causal Reasoning, Natural Language Inference, Abstractive Summarization, and Machine Translation, several of which were previously unavailable for Burmese. The benchmark is constructed through a rigorous native-speaker-driven process to ensure linguistic naturalness, fluency, and cultural authenticity while minimizing translation-induced artifacts. We conduct a large-scale evaluation of both open-weight and commercial LLMs to examine challenges in Burmese modeling arising from limited pretraining coverage, rich morphology, and syntactic variation. Our results show that Burmese performance depends more on architectural design, language representation, and instruction tuning than on model scale alone. In particular, Southeast Asia regional fine-tuning and newer model generations yield substantial gains. Finally, we release BURMESE-SAN as a public leaderboard to support systematic evaluation and sustained progress in Burmese and other low-resource languages. https://leaderboard.sea-lion.ai/detailed/MY

Masked Diffusion Models (MDMs) enable flexible decoding orders, yet existing samplers remain largely greedy, selecting locally certain tokens without accounting for their downstream effects. We show that this myopia can increase cumulative uncertainty and lead to suboptimal generation. To address this, we propose the **Info-Gain Sampler**, a training-free decoding method that uses the bidirectional structure of MDMs to balance immediate uncertainty with the information gained over remaining masked positions. Across reasoning, coding, creative writing, and image generation tasks, Info-Gain Sampler consistently outperforms existing MDM samplers, improving average reasoning accuracy by 2.9--11.6 percentage points and achieving a 62.8% average win rate in creative writing. The code is available at https://github.com/yks23/Information-Gain-Sampler.

Recent work has shown that language models (LMs) trained on synthetic corpora can exhibit typological preferences that resemble cross-linguistic regularities in human languages, particularly for syntactic phenomena such as word order. In this paper, we extend this paradigm to differential argument marking (DAM), a semantic licensing system in which morphological marking depends on semantic prominence. Using a controlled synthetic learning method, we train GPT-2 models on 18 corpora implementing distinct DAM systems and evaluate their generalization using minimal pairs. Our results reveal a dissociation between two typological dimensions of DAM. Models reliably exhibit human-like preferences for natural markedness direction, favoring systems in which overt marking targets semantically atypical arguments. In contrast, models do not reproduce the strong object preference in human languages, in which overt marking in DAM more often targets objects rather than subjects. These findings suggest that different typological tendencies may arise from distinct underlying sources.

Frontier AI systems are increasingly capable and deployed in high-stakes multi-agent environments. However, existing AI safety benchmarks largely evaluate single agents, leaving multi-agent risks such as coordination failure and conflict poorly understood. We introduce GT-HarmBench, a benchmark of 1,535 high-stakes scenarios spanning game-theoretic structures such as the Prisoner's Dilemma, Stag Hunt and Chicken. Scenarios are drawn from realistic AI risk contexts in the MIT AI Risk Repository. Across 15 frontier models, agents fail to choose socially beneficial actions in 38% of high-stakes cases, such as military escalation, election manipulation, and medical malpractice. We measure sensitivity to game-theoretic prompt framing and ordering, and analyze reasoning patterns driving failures. We further show that game-theoretic interventions improve socially beneficial outcomes by up to 18%. Our results highlight substantial reliability gaps and provide a broad standardized testbed for studying alignment in multi-agent environments. The benchmark and code are available at https://github.com/causalNLP/gt-harmbench.

Modern LLM-based agents and chat assistants rely on long-term memory frameworks to store reusable knowledge, recall user preferences, and augment reasoning. As researchers create more complex memory architectures, it becomes increasingly difficult to analyze their capabilities and guide future memory designs. Most long-term memory benchmarks focus on simple fact retention, multi-hop recall, and time-based changes. While undoubtedly important, these capabilities can often be achieved with simple retrieval-augmented LLMs and do not test complex memory hierarchies. To bridge this gap, we propose StructMemEval - a benchmark that tests the agent's ability to organize its long-term memory, not just factual recall. We gather a suite of tasks that humans solve by organizing their knowledge in a specific structure: transaction ledgers, to-do lists, trees and others. Our initial experiments show that simple retrieval-augmented LLMs struggle with these tasks, whereas memory agents can reliably solve them if prompted how to organize their memory. However, we also find that modern LLMs do not always recognize the memory structure when not prompted to do so. This highlights an important direction for future improvements in both LLM training and memory frameworks.

Diffusion large language models (dLLMs) have recently gained significant attention due to their inherent support for parallel decoding. Building on this paradigm, Mixture-of-Experts (MoE) dLLMs with autoregressive (AR) initialization have further demonstrated strong performance competitive with mainstream AR models. However, we identify a fundamental mismatch between MoE architectures and diffusion-based decoding. Specifically, a large number of experts are activated at each denoising step, while only a small subset of tokens is ultimately accepted, resulting in substantial inference overhead and limiting their deployment in latency-sensitive applications. In this work, we propose TEAM, a plug-and-play framework that accelerates MoE dLLMs by enabling more accepted tokens with fewer activated experts. TEAM is motivated by the observation that expert routing decisions exhibit strong temporal consistency across denoising levels as well as spatial consistency across token positions. Leveraging these properties, TEAM employs three complementary expert activation and decoding strategies, conservatively selecting necessary experts for decoded and masked tokens and simultaneously performing aggressive speculative exploration across multiple candidates. Experimental results demonstrate that TEAM achieves up to 2.2x speedup over vanilla MoE dLLM, with negligible performance degradation. Code is released at https://github.com/PKU-SEC-Lab/TEAM-MoE-dLLM.

Large language models (LLMs) are increasingly deployed as educational agents for automatic short answer grading (ASAG) in real-world educational environments, significantly boosting assessment efficiency and scalability. However, when these grading agents operate ``in the wild'', their vulnerability to adversarial manipulation raises critical concerns about agent security and trustworthiness. In this paper, we introduce GradingAttack, a fine-grained adversarial attack framework that systematically evaluates the security vulnerabilities of LLM based educational grading agents. Specifically, we design token-level and prompt-level attack strategies that manipulate agent grading outcomes while maintaining high stealth, exposing fundamental weaknesses in current agent deployments. Experiments on multiple datasets demonstrate that both attack strategies effectively compromise grading agents, with prompt-level attacks achieving higher success rates and token-level attacks exhibiting superior stealth capability. Our findings reveal that current LLM based educational agents lack robust defenses against adversarial attacks, underscoring the urgent need for developing secure and trustworthy agent systems for critical educational applications.

Transformers are arguably the preferred architecture for language generation. In this paper, inspired by continued fractions, we introduce a new function class for generative modeling. The architecture family implementing this function class is named CoFrGeNets - Continued Fraction Generative Networks. We design novel architectural components based on this function class that can replace Multi-head Attention and Feed-Forward Networks in Transformer blocks while requiring much fewer parameters. We derive custom gradient formulations to optimize the proposed components more accurately and efficiently than using standard PyTorch-based gradients. Our components are a plug-in replacement requiring little change in training or inference procedures that have already been put in place for Transformer-based models thus making our approach easy to incorporate in large industrial workflows. We experiment on two very different transformer architectures GPT2-xl (1.5B) and Llama3 (3.2B), where the former we pre-train on OpenWebText and GneissWeb, while the latter we pre-train on the docling data mix which consists of nine different datasets. Results show that the performance on downstream classification, Q\& A, reasoning and text understanding tasks of our models is competitive and sometimes even superior to the original models with $\frac{2}{3}$ to $\frac{1}{2}$ the parameters and shorter pre-training time. We believe that future implementations customized to hardware will further bring out the true potential of our architectures.

Estimating task progress requires reasoning over long-horizon dynamics rather than recognizing static visual content. While modern Vision-Language Models (VLMs) excel at describing what is visible, it remains unclear whether they can infer how far a task has progressed from partial observations. To this end, we introduce Progress-Bench, a benchmark for systematically evaluating progress reasoning in VLMs. Beyond benchmarking, we further explore a human-inspired two-stage progress reasoning paradigm through both training-free prompting and training-based approach based on curated dataset ProgressLM-45K. Experiments on 14 VLMs show that most models are not yet ready for task progress estimation, exhibiting sensitivity to demonstration modality and viewpoint changes, as well as poor handling of unanswerable cases. While training-free prompting that enforces structured progress reasoning yields limited and model-dependent gains, the training-based ProgressLM-3B achieves consistent improvements even at a small model scale, despite being trained on a task set fully disjoint from the evaluation tasks. Further analyses reveal characteristic error patterns and clarify when and why progress reasoning succeeds or fails. Website: https://progresslm.github.io/ProgressLM/

AI agents have seen widespread adoption in information retrieval for scientific research, giving rise to tools such as Deep Research. However, existing retrieval agents mainly rely on keyword- or embedding-based methods. While effective at capturing content-level similarities, they struggle to understand complex relational networks among scientific papers, such as identifying corroborating or conflicting studies and tracing technological lineages. This fundamental limitation often results in fragmented knowledge structures, misinterpreted research sentiment, and ineffective modeling of collective scientific progress. To address this limitation, we introduce SciNet, the first Scientific Network relation-aware dataset for information retrieval agents. Built on a meta-database of 269 million papers across 7 disciplines and containing 8,940 carefully designed tasks, SciNet systematically captures three levels of relational understanding: ego-centric retrieval of papers with novel knowledge structures, pairwise identification of scholarly relationships, and path-wise reconstruction of scientific evolution. Extensive evaluation of three categories of retrieval agents shows that their accuracy on relation-aware tasks often falls below 20%, highlighting a fundamental shortcoming of current retrieval paradigms. Importantly, in a downstream literature review application, agents empowered with SciNet achieve a 25.3% improvement in review quality, highlighting the critical value of relation-aware retrieval for deepening scientific insights. We publicly release SciNet at https://github.com/tsinghua-fib-lab/SciNet to support future research.

Growing reliance on LLMs for psychiatric self-assessment raises questions about their ability to interpret qualitative patient narratives. This depth over breadth case study directly compares state-of-the-art LLMs and mental health professionals in assessing Borderline (BPD) and Narcissistic (NPD) Personality Disorders based on Polish-language first-person autobiographical accounts. Within our sample, the overall diagnostic scores of the top-performing Gemini Pro models (65.48%) were 21.91 percentage points higher than the average scores of the human professionals (43.57%). While both models and human experts excelled at identifying BPD (F1 = 83.4 & F1 = 80.0, respectively), models severely underdiagnosed NPD (F1 = 6.7 vs. 50.0), showing a potential reluctance toward the value-laden term "narcissism." Qualitatively, models provided confident, elaborate justifications focused on patterns and formal categories, while human experts remained concise and cautious, emphasizing the patients' sense of self and temporal experience. Our findings demonstrate that while LLMs might be competent at interpreting complex first-person clinical data, their outputs still carry critical reliability and bias issues.

Scaling the context length of large language models (LLMs) offers significant benefits but is computationally expensive. This expense stems primarily from the self-attention mechanism, whose $O(N^2)$ complexity with respect to sequence length presents a major bottleneck for both memory and latency. Fortunately, the attention matrix is often sparse, particularly for long sequences, suggesting an opportunity for optimization. Block-sparse attention has emerged as a promising solution that partitions sequences into blocks and skips computation for a subset of these blocks. However, the effectiveness of this method is highly dependent on the underlying attention patterns, which can lead to sub-optimal block-level sparsity. For instance, important key tokens for queries within a single block may be scattered across numerous other blocks, leading to computational redundancy. In this work, we propose Permuted Block-Sparse Attention (\textbf{PBS-Attn}), a plug-and-play method that leverages the permutation properties of attention to increase block-level sparsity and enhance the computational efficiency of LLM prefilling. We conduct comprehensive experiments on challenging real-world long-context datasets, demonstrating that PBS-Attn consistently outperforms existing block-sparse attention methods in model accuracy and closely matches the full attention baseline. Powered by our custom permuted-FlashAttention kernels, PBS-Attn achieves an end-to-end speedup of up to $2.75\times$ in long-context prefilling, confirming its practical viability. Code available at https://github.com/xinghaow99/pbs-attn

Supervised fine-tuning (SFT) is the standard approach for post-training large language models (LLMs), yet it often shows limited generalization. We trace this limitation to its default training objective: negative log likelihood (NLL). While NLL is classically optimal when training from scratch, post-training operates in a different paradigm and could violate its optimality assumptions, where models already encode task-relevant priors and supervision can be long and noisy. In this work, we systematically study various probability-based objectives and characterize when and why different objectives succeed or fail under varying conditions. Through comprehensive experiments and extensive ablation studies across 8 model backbones, 27 benchmarks, and 7 domains, we uncover a critical dimension that governs objective behavior: the model-capability continuum. Near the model-strong end, prior-leaning objectives that downweight low-probability tokens (e.g., $-p$, $-p^{10}$, thresholded variants) consistently outperform NLL; toward the model-weak end, NLL dominates; in between, no single objective prevails. Our theoretical analysis further elucidates how objectives trade places across the continuum, providing a principled foundation for adapting objectives to model capability. The code is available at https://github.com/GaotangLi/Beyond-Log-Likelihood.

Knowledge-graph retrieval-augmented generation (KG-RAG) couples large language models (LLMs) with structured, verifiable knowledge graphs (KGs) to reduce hallucination and provide reasoning traces. However, current KG-RAG systems often rely on fixed pipelines of multiple LLM modules (e.g., planning, reasoning, and responding), which inflate inference costs and tie performance to specific graph schemas. To address this, we introduce KG-R1, an agentic framework that optimizes KG-RAG through reinforcement learning (RL). Unlike modular workflows, KG-R1 uses a single agent that interacts with KGs as its environment, learning to retrieve information at each step and incorporating it into its reasoning and generation in a unified process. Across Knowledge-Graph Question Answering (KGQA) benchmarks, KG-R1 demonstrates both efficiency and transferability-using Qwen 2.5-3B, KG-R1 improves answer accuracy with fewer generation tokens than prior multi-module workflow methods that use much larger foundation or fine-tuned models. Furthermore, KG-R1 exhibits strong plug-and-play capability: after training, maintaining accuracy on unseen KGs without retraining. These properties make KG-R1 a promising KG-RAG framework for real-world deployment. Our code is publicly available at github.com/junhongmit/KG-R1/.

Building interactive omni-modal assistants often relies on end-to-end multimodal alignment to fuse heterogeneous modalities, which incurs substantial data and compute costs and limits extensibility. We present Training-Free Large Language Model Orchestration (LLM Orchestration), a training-free orchestration framework that integrates off-the-shelf modality experts into a unified multimodal input--output system without additional gradient-based training for integration. LLM Orchestration comprises three components: (1) an LLM controller that infers user intent and emits explicit control tokens for expert selection and sequencing, enabling protocol-constrained and auditable routing; (2) a text-centric cross-modal memory that compresses multimodal evidence into structured records for lightweight retrieval and reuse, reducing redundant expert invocations across turns; and (3) a unified interaction layer that executes routing and memory decisions to support consistent modality transitions, full-duplex streaming, and interruption-aware dialogue. Across diverse multimodal benchmarks, LLM Orchestration achieves strong performance under standard evaluation constraints while maintaining low orchestration overhead and modular upgradeability, providing a practical alternative to costly joint training for omni-modal systems.

Despite advances in legal NLP, no comprehensive evaluation of Transformer-based models customized for legal tasks (referred to as `legal-specific' models in this paper) exists for contract classification tasks. To address this gap, we present an evaluation of 13 legal-specific transformer-based models on 3 English-language contract classification tasks and compare them with 9 generalist models. The results show that legal-specific models consistently outperform generalist models, especially on tasks requiring nuanced legal understanding. They also help reduce misclassification of rare classes in imbalanced datasets. Legal-BERT and Contracts-BERT establish new SOTAs on two of the three tasks, despite having 69% fewer parameters than the best-performing generalist models. We also identify CaseLaw-BERT and LexLM as strong additional baselines for contract classification. Our results highlight the shortcomings of generalist models, emphasizing the need for domain-specific customization, particularly in the context of legal applications.

Multimodal foundation models have demonstrated impressive capabilities across diverse tasks. However, their potential as plug-and-play solutions for missing modality reconstruction remains underexplored. To bridge this gap, we identify and formalize three potential paradigms for missing modality reconstruction, and perform a comprehensive evaluation across these paradigms, covering 42 model variants in terms of reconstruction accuracy and adaptability to downstream tasks. Our analysis reveals that current foundation models often fall short in two critical aspects: (i) fine-grained semantic extraction from the available modalities, and (ii) robust validation of generated modalities. These limitations lead to suboptimal and, at times, misaligned generations. To address these challenges, we propose an agentic framework tailored for missing modality reconstruction. This framework dynamically formulates modality-aware mining strategies based on the input context, facilitating the extraction of richer and more discriminative semantic features. In addition, we introduce a self-refinement mechanism, which iteratively verifies and enhances the quality of generated modalities through internal feedback. Experimental results show that our method reduces FID for missing image reconstruction by at least 14\% and MER for missing text reconstruction by at least 10\% compared to baselines. Code are released at: https://github.com/Guanzhou-Ke/AFM2.

Multi-modal large language models (MLLMs) have rapidly advanced in visual tasks, yet their spatial understanding remains limited to single images, leaving them ill-suited for physical-world applications that require multi-frame reasoning. In this paper, we propose a framework to equip MLLMs with multi-frame spatial understanding by integrating fundamental spatial skills, including depth perception, visual correspondence, and dynamic perception. We design a novel data pipeline and collect the MultiSPA dataset of more than 27 million samples spanning diverse 3D and 4D scenes to enable training. Alongside MultiSPA, we introduce a comprehensive benchmark that tests a wide spectrum of spatial tasks under uniform metrics. Our resulting model, Multi-SpatialMLLM, achieves significant gains over baselines and proprietary systems, demonstrating scalable and generalizable multi-frame perception. We further observe multi-task benefits and emergent spatial capabilities in challenging scenarios, and showcase how our model can serve as a multi-frame reward annotator for robotics.

Recent advances in large language models (LLMs) have intensified efforts to fuse heterogeneous open-source models into a unified system that inherits their complementary strengths. Existing logit-based fusion methods maintain inference efficiency but treat vocabulary dimensions independently, overlooking semantic dependencies encoded by cross-dimension interactions. These dependencies reflect how token types interact under a model's internal reasoning and are essential for aligning models with diverse generation behaviors. To explicitly model these dependencies, we propose \textbf{InfiGFusion}, the first structure-aware fusion framework with a novel \textit{Graph-on-Logits Distillation} (GLD) loss. Specifically, we retain the top-$k$ logits per output and aggregate their outer products across sequence positions to form a global co-activation graph, where nodes represent vocabulary channels and edges quantify their joint activations. To ensure scalability and efficiency, we design a sorting-based closed-form approximation that reduces the original $O(n^4)$ cost of Gromov-Wasserstein distance to $O(n \log n)$, with provable approximation guarantees. Experiments across multiple fusion settings show that GLD consistently improves fusion quality and stability. InfiGFusion outperforms SOTA models and fusion baselines across 11 benchmarks spanning reasoning, coding, and mathematics. It shows particular strength in complex reasoning tasks, with +35.6 improvement on Multistep Arithmetic and +37.06 on Causal Judgement over SFT, demonstrating superior multi-step and relational inference.

Pretraining data curation is a cornerstone in Large Language Model (LLM) development, leading to growing research on quality filtering of large web corpora. From statistical quality flags to LLM-based labelling systems, datasets are divided into categories, frequently reducing to a binary: those passing the filters are deemed as valuable examples, others are discarded as useless or detrimental. However, a more detailed understanding of the contribution of different kinds of texts to model performance is still largely lacking. In this article, we present the first study utilising registers or genres - a widely used standard in corpus linguistics to model linguistic variation - to curate pretraining datasets and investigate the effect of register on the performance of LLMs. We train small generative models with register classified data and evaluate them using standard benchmarks, and show that the register of pretraining data substantially affects model performance. We uncover surprising relationships between the pretraining material and the resulting models: using the News register results in subpar performance, and on the contrary, including the Opinion class, covering texts such as reviews and opinion blogs, is highly beneficial. While a model trained on the entire unfiltered dataset outperforms those trained on datasets limited to a single register, combining well-performing registers like How-to-Instructions, Informational Description, and Opinion leads to major improvements. Furthermore, analysis of individual benchmark results reveals key differences in the strengths and drawbacks of specific register classes as pretraining data. These findings show that register is an important explainer of model variation and can facilitate more deliberate future data selection practices.

Dense vector retrieval is an important building block of modern machine learning systems, underlying applications ranging from semantic search to retrieval-augmented generation and knowledge-intensive reasoning. Beyond retrieving items that are individually similar to a query, many applications require a set of results that is also diverse, complementary, and collectively informative. Balancing similarity and diversity is therefore central to effective retrieval, but remains challenging to optimize in a stable and theoretically grounded way. Maximal Marginal Relevance (MMR) is a widely adopted heuristic for this problem, yet its reliance on a manually tuned parameter leads to optimization fluctuations and unpredictable retrieval results. More broadly, existing methods provide limited theoretical insight into how similarity and diversity interact in dense vector spaces, leaving the joint optimization problem insufficiently understood. To address these challenges, this paper introduces a novel approach that characterizes both constraints simultaneously by maximizing the similarity between the query vector and the sum of the selected candidate vectors. We formally define this optimization problem, Vector Retrieval with Similarity and Diversity (VRSD), and prove that it is NP-complete, establishing a rigorous theoretical bound on the inherent difficulty of this dual-objective retrieval. Subsequently, we present a parameter-free heuristic algorithm to solve VRSD. Extensive evaluations on multiple datasets, incorporating both objective geometric metrics and LLM-simulated subjective assessments, demonstrate that our VRSD heuristic consistently outperforms established baselines, including MMR and Determinantal Point Processes (k-DPP).

We aim to improve the reasoning capabilities of diffusion language models (DLMs). While SFT is a popular post-training recipe for autoregressive models, its use in DLMs faces challenges and can even hurt performance, though the underlying causes remain understudied. Our analysis reveals that vanilla SFT overlooks learnability, namely what and when tokens are learned. Specifically, rare tokens are difficult to learn when most of the input is masked, whereas it is straightforward and thus of little value to learn common tokens when most of the input is unmasked. Motivated by our analysis, we propose LIFT, an efficient SFT-based post-training algorithm for DLMs. LIFT learns easy tokens when most of the input is masked and hard tokens when more context is available, thus aligning the training with the information available at different diffusion time steps. Our results show that LIFT outperforms existing SFT baselines across six reasoning benchmarks, achieving up to a 3x relative gain on AIME'24 and AIME'25. Our code is publicly available at https://github.com/divelab/LIFT.

Inference-time scaling can reduce errors in structured query generation, but methods to allocate the compute for query code generation remains underexplored. We study Text2Cypher, where language models generate Cypher queries that execute against property graph databases. Non-executable queries constitute a distinct syntactic failure separate from semantic inaccuracy: a syntax error triggers a system-generated error message from the database. These error messages are typically discarded at inference time rather than leveraged through in-context learning (ICL). We compare two inference methods: Independent Scaling (IS), which performs memoryless resampling, and Reflection-Augmented Scaling (RAS), which conditions each new attempt on prior execution feedback via ICL. Across three Neo4j datasets and five code-specialized language models, RAS reduces the Query Execution Error Rate by 41--50% at n{=}5, outperforming IS at 32--38%. Execution errors are not merely failures to discard but actionable feedback, and structuring inference-time compute around them is a more efficient path to executability than scaling independent samples.

Large language models can answer questions about textbooks, lecture notes, and programming exercises more reliably when their answers are grounded in an explicit knowledge source. Retrieval-augmented generation (RAG) is a common approach: relevant fragments of a document are retrieved and inserted into the model context before answering. For mathematical and technical material, the original LaTeX source can be a better starting point than a PDF, because it contains structural information, labels, sectioning commands, macros, and authorial intent that are often lost or distorted in PDF extraction. However, LaTeX source is not automatically AI-friendly. Cross-references must be resolved, custom macros must be interpreted, exercises and examples must be identified, and author-supplied semantic metadata may be needed. This article describes a focused preprocessing approach for turning LaTeX source, together with its compiled auxiliary files and optional author annotations, into Markdown and JSONL chunks suitable for indexing in a vector database.

Self-evolving agents should not train on examples they cannot justify. Data-free self-evolving search agents offer a scalable route to systems that generate their own questions, answer them, and improve from their own feedback without human annotations. Yet, without verifiable evidence, this loop can reward fluent but unsupported examples, turning the self-generated curriculum into an opaque and potentially unreliable training signal. We argue that evidence verifiability is a prerequisite for trustworthy self-evolution in search agents: each generated instance should include not only an answer but also a source-grounded span whose contribution to that answer can be measured. We introduce EVE-Agent, an Evidence-Verifiable Self-Evolving Agent that operationalizes this principle through a modification to the proposer--solver framework. The proposer generates a question, an answer, and a verbatim evidence span. An evidence verifier then rewards the span according to the marginal accuracy gain when the evidence is provided. This produces a training signal that favors evidence that genuinely helps answer the question, without requiring oracle answers, human labels, or external annotations. EVE-Agent leaves the backbone model, retriever, search tool, and optimization framework unchanged. Experiments show that EVE-Agent substantially improves evidence-grounded correctness over prior self-evolving search agents. The resulting curriculum is not merely self-generated but auditable by construction: each training example carries an inspectable source span that explains why it should be trusted.

Benchmark accuracy is often implicitly assumed to reflect grounded visual understanding in vision-language models (VLMs), yet it remains unclear to what extent such scores truly reflect reliance on visual evidence. Motivated by a surprising observation that removing a substantial fraction of image tokens only degrades model performance very slightly on a widely used hallucination benchmark, we systematically investigate this mismatch in a set of open-source VLMs. Our analysis spans multiple levels of granularity, spanning global visual degradation, localized occlusion, question reformulation, answer-space expansion, and decision-level analyses beyond standard accuracy. We further complement these behavioral results with a layer-wise analysis of vision-token geometry. Throughout the experiments, we find that although VLMs do incorporate visual input, their predictions are less sensitive to the loss of fine-grained visual evidence that standard accuracy should have suggested. Even when the final prediction remains unchanged, the model's internal support for the correct answer may already be weakened. We further complement a representation-level analysis, which shows increasing similarity among visual tokens in deeper layers, providing a possible explanation for our findings. Together, these results suggest that current benchmarks are not sufficient to reliably evaluate fine-grained visual grounding in VLMs.

Generative Vision-Language Models (VLMs) perform well on multimodal reasoning, but how visual inputs are transformed to text remains poorly understood. Existing interpretability work on VLMs uses Sparse Autoencoders (SAEs), which decompose static residual representations and miss the functional updates that drive cross-modal interaction. We adopt a function-centric framework based on Transcoders, sparse approximations of MLP sublayers that act as a causal proxy for layer-wise computation. Applied to Gemma 3-4B-IT, the framework decomposes the model into interpretable computational pathways linking image patches to directions in token generation. Transcoder attributions produce stronger and more stable effects on visually grounded tokens under patch ablation than SAE attributions, and align better with semantically relevant image regions. A False Visual Grounding counterfactual analysis confirms that the recovered pathways are specific to vision-language interaction.Finally, we perform a structural analysis of hallucinated generations, by extracting graph-based indicators from circuit traces produced by the transcoders. A logistic classifier over these mechanistic graph features predicts hallucinations at AUC $0.68$. These results show that function-centric circuit decomposition yields interpretable and predictive accounts of multimodal computation in VLMs.

Formal mathematics libraries are rapidly expanding, creating a growing need to refactor verified proofs for maintainability and to improve training data quality for neural provers. However, scalable proof optimization is hindered by heterogeneous and heuristically specified objectives, scarce data, and high training and inference costs. To overcome these challenges, we introduce ImProver 2, a neurosymbolic framework for automated proof optimization in Lean 4. ImProver 2 combines a data-efficient expert-iteration pipeline with a scaffold that exposes formal structure alongside lightweight informal abstractions. We further introduce a suite of metrics capturing structural proof properties. Using ImProver 2, we train a 7B-parameter model that outperforms orders-of-magnitude larger models within the same model family, and is competitive with mid-tier frontier models across metrics. We additionally demonstrate that our neurosymbolic scaffold significantly improves performance across both small and frontier models. We show that with proper scaffolding and training, small models can effectively restructure research-level proofs over complex and varied metrics, matching substantially larger systems and establishing proof optimization as a scalable, learnable task.

As large language model (LLM)-based agents increasingly participate in online discourse, red-teaming their capacity to support political influence campaigns is critical for information integrity. In pursuit of this goal, we focus on locally deployed open-source LLMs, as opposed to frontier API-only models, given their superior alignment with the operational constraints of privacy-conscious malicious actors deployed in social media environments. We introduce an empirical red-teaming framework for measuring LLM Overton Windows (OWs), defined as the range of political opinions a model can reliably express on controversial topics, and for quantifying how simple natural-language jailbreaks expand that range. We evaluate more than 30 LLMs spanning 10 model families and five countries of origin. We find systematic asymmetries in political expressivity: open-source LLMs are typically more willing to generate left-leaning social media content, OWs tend to contract inversely to model size, and regional differences are substantial despite uneven representation in the open-source ecosystem. Jailbreak potency also varies sharply across model families, motivating a workflow for identifying effective combinations of jailbreak techniques. Taken together, our results establish a practical framework for auditing the political steerability of open-source LLMs and for helping future researchers design stronger countermeasures against LLM-enabled influence campaigns.

The exponential growth of global academic output has confronted researchers and AI agents with an unprecedented ``information explosion,'' where fragmented and unstructured knowledge organization impedes deep interdisciplinary integration. Current academic retrieval tools predominantly rely on superficial keyword matching or vector-space semantic retrieval, which lack the topological reasoning capabilities required to navigate complex logical connections. Agentic deep-research-based frameworks are often prone to logical hallucinations and consuming high inference costs. To bridge this gap, in this report, we introduce SciAtlas, a large-scale, multi-disciplinary, heterogeneous academic resource knowledge graph designed as a panoramic scientific evolution network. By integrating over 43M papers from 26 disciplines, and a total of 157M entities and 3B triplets, SciAtlas provides a structured topological cognitive substrate that dismantles disciplinary barriers and furnishes AI agents with a global perspective. Furthermore, we develop a neuro-symbolic retrieval algorithm featuring tri-path collaborative recall and graph reranking, achieving a seamless transition from simple semantic matching to deterministic association discovery. We also present key application directions of SciAtlas, including literature review, automated research trend synthesis, idea positioning, and academic trajectory exploration, to demonstrate that SciAtlas can serve as an effective ``cognitive map'' to empower the full loop of automated scientific research while significantly reducing reasoning costs. We have released the interfaces for KG retrieval and various downstream tasks in our GitHub repo.

Chain-of-thought (CoT) prompting is necessary for arithmetic in small language models, yet shuffling its steps preserves most performance. What does CoT contribute if not logical sequencing? In three 1-3B instruction-tuned LMs on GSM8K, we isolate the answer-readout stage via prefix completion and identify a positional shortcut: the model copies whichever number occupies the trailing position before the answer delimiter, regardless of intermediate reasoning. Gold-answer presence accounts for 54-92 pp of accuracy (89-92% of each model's teacher-forcing ceiling); even on incorrect items, the final answer matches the last CoT number 95-96% of the time. The copy channel takes precedence over retained-context completion: replacing the trailing number with a wrong value collapses accuracy to near-zero despite correct intermediates, yet removing it recovers 5-32 pp above that floor--even single-step arithmetic the model can otherwise perform is suppressed when a copyable number is present. Qwen and Llama copy novel distractors 87-95% of the time; Gemma gates selectively. Head-level ablation implicates architecture-specific head sets; the effect replicates on GSM-Symbolic. On non-arithmetic BBH tasks, shuffle retention drops sharply; at 7-8B, content-selective gating emerges. Step-level faithfulness evaluations risk conflating positional answer transport with genuine computation--a failure mode for CoT-based oversight.

Personalized pricing negotiations are a challenging testbed for LLM agents because successful interaction does not guarantee profitable decision making. A seller may produce valid actions and close many deals while still pricing poorly when buyer willingness to pay and bargaining traits remain hidden. This paper presents PrefBench, a simulator-based benchmark for hidden-preference personalized pricing negotiations. Each episode pairs a simulated buyer with a fixed vehicle-customization bundle; the seller observes public persona descriptors, bundle information, and negotiation history, while latent buyer variables govern valuation, patience, counter-offer behavior, and walkaway decisions. PrefBench evaluates this setting through an LLM-facing state-summary protocol that constrains agents to return strict JSON actions under a fixed hidden-information boundary. We evaluate zero-shot LLM sellers against heuristic references over 7,500 episodes. The tested LLMs follow the protocol reliably and achieve deal rates above 0.99, but their seller-profit outcomes remain weak: the best LLM average profit is only slightly above the random baseline and far below a simple concession heuristic under the same episode stream. These results show that structured action compliance and agreement-seeking behavior can coexist with weak profit-sensitive bargaining. PrefBench provides a controlled benchmark for evaluating pricing-agent behavior under hidden buyer preferences.

Text-to-SQL converts natural language questions into executable SQL queries, enabling non-technical users to access relational databases for analytics and intelligent data services. In real-world scenarios, performance is often constrained by low-resource settings, where high-quality annotated \texttt{<question, SQL>} pairs are scarce, particularly for domain-specific databases. Additional challenges include opaque schema definitions, abbreviations, and implicit business logic that are not explicitly encoded in the schema. Existing data synthesis and prompting techniques improve coverage but often fail to produce task-specific, semantically grounded examples aligned with database constraints. To address these challenges, we propose a knowledge-aware Text-to-SQL framework that constructs task-specific knowledge base including schema semantics, abbreviations, business logic, and query patterns, and injects them into both training and inference. This framework generates diverse, contextually grounded synthetic training data and enhances inference through targeted knowledge retrieval. Experiments on seven benchmarks, covering both general and domain-specific datasets, demonstrate that our approach substantially improves the performance of open-source and closed-source large language models in Text-to-SQL tasks, especially in low-resource domain-specific settings, enhancing generalization, robustness, and adaptability.

What happens when the strongest alliance member pressures a weaker member over territory and strategic control? We examine the Greenland sovereignty crisis as a stress test for LLM geopolitics, centered on the 2019-2026 U.S. push to acquire Greenland from the Kingdom of Denmark. The crisis nests two collective-action problems: Arctic strategic control and whether NATO can enforce alliance norms against the dominant member. We develop three games (asymmetric coercion; a NATO assurance game with a critical-mass tipping point; a triadic extensive-form game with social preferences) and test them with a multi-agent simulation in which eight frontier LLMs play six geopolitical roles (United States, Denmark, Greenland, NATO, Russia, Canada) across 3,604 completed games and 108,120 action observations. Using inverse game theory, we recover each model's structural utility parameters (alpha, beta, gamma, delta, eta) for material self-interest, reciprocity, inequality aversion, norm respect, and commitment consistency. Three findings stand out. First, all eight models become more escalatory under coercion framing (four-action escalation rises from 10.7% to 28.6%). Second, Chinese-origin models show systematically different power-weight profiles from Western-origin models when playing the U.S. role. Third, peaceful US acquisition emerges in only 1.9% of clean games and only 3 of 8 frontier models ever achieve it, most prominently DeepSeek V3.2, which executes a stable five-round playbook through the metropole. Prompts emphasizing jus cogens and self-determination reduce escalation back near baseline in the English-only confirmatory sample; multilingual contrasts are reported as exploratory sensitivity checks. We position this as a structural benchmark for LLM geopolitical behavior, complementing action-frequency benchmarks.

This survey provides a comprehensive catalog of publicly available text and speech resources for two West African languages: Hausa, an Afroasiatic language with approximately 80-100 million speakers, and Fongbe, a Niger-Congo language spoken by approximately 2 million people in Benin. These languages represent contrasting cases on the resource availability spectrum. We address the question: \textit{What is the current state of publicly available NLP resources for Hausa and Fongbe, and what gaps remain?} Through systematic search of academic repositories, data platforms, and web sources, we catalog parallel corpora, monolingual text collections, speech datasets, pre-trained models, and evaluation benchmarks. For each resource, we document size, domain coverage, format, licensing, and accessibility. Our findings reveal that Hausa benefits from broader text resource diversity across news, encyclopedic, and educational domains. Fongbe, while having more limited text resources, has been the focus of recent academic speech data collection initiatives. Both languages are represented in Masakhane benchmarks for NER and POS tagging. We provide task-specific recommendations and identify priority gaps including domain-diverse Fongbe text and dedicated Hausa speech corpora.

Quantifying the deceptive potential of Large Language Models (LLMs) is critical for AI safety, yet difficult to achieve in uncontrolled environments. This work investigates the reasoning, persuasion, and deceptive capabilities of LLMs within the social deduction game Secret Hitler. I introduce an open-source framework and novel metrics to measure performance: Role Identification Accuracy, Deception Retention Rate, and Game State Impact Rate. By benchmarking models against rule-based algorithms and human games, I identify a gap between conversational ability and strategic depth. The study also analyzes the impact of reasoning-enhancement techniques on win rates and strategic reasoning. Neither Chain-of-Thought prompting nor internal memory bring improvements in performance, with up to 23.2% worse win rates for fascist roles. While rule-based agents align with expert human voting decisions 86.7% of the time, models like Llama 3.1 70B achieve only a 59.7% accuracy. Models playing as Fascists consistently yield negative impact scores and fail to sustain deception, resulting in roughly 40% shorter games compared to humans. These findings suggest that current architectures remain ineffective at complex, multi-turn manipulation. As capabilities advance, detecting when models begin to master these deceptive behaviors is crucial. The developed framework serves as a reproducible testbed for future alignment research.

Selective layer-wise updates are essential for low-cost continued pre-training of Large Language Models (LLMs), yet determining which layers to freeze or train remains an empirical black-box problem due to the lack of interpretable guidance. To address this issue, we propose LayerTracer, an architecture-agnostic diagnostic framework that reveals the evolution patterns of layer-wise representations and stability by locating task execution positions and quantifying layer sensitivity. Analysis results reveal that deep layers act as critical regions for task execution and maintain high stability against disruptive updates. Guided by this finding, we conduct three controlled continued pre-training trials to compare diverse freeze-train strategies, demonstrating that training shallow layers while freezing deep layers consistently outperforms full-parameter fine-tuning and the opposite allocation on both C-Eval and CMMLU benchmarks. We further present a hybrid model case study, which validates that placing high-quality pre-trained modules in deep layers effectively preserves inherent knowledge of the model. This work delivers a low-cost and interpretable solution for resource-constrained teams, offering actionable guidance for layer-wise parameter allocation in continued pre-training and hybrid model construction.

While multi-agent systems (MAS) promise elevated intelligence through coordination of agents, current approaches to automatic MAS design under-deliver. Such shortcomings stem from two key factors: (1) methodological complexity - agent orchestration is performed using sequential, code-level execution that limits global system-level holistic reasoning and scales poorly with agent complexity - and (2) efficacy uncertainty - MAS are deployed without understanding if there are tangible benefits compared to single-agent systems (SAS). We propose MASOrchestra, a training-time framework that formulates MAS orchestration as a function-calling reinforcement learning problem with holistic orchestration, generating an entire MAS at once. In MAS-Orchestra, complex, goal-oriented subagents are abstracted as callable functions, enabling global reasoning over system structure while hiding internal execution details. To rigorously study when and why MAS are beneficial, we introduce MASBENCH, a controlled benchmark that characterizes tasks along five axes: Depth, Horizon, Breadth, Parallel, and Robustness. Our analysis reveals that MAS gains depend critically on task structure, verification protocols, and the capabilities of both orchestrator and subagents, rather than holding universally. Guided by these insights, MAS-Orchestra achieves consistent improvements on public benchmarks including mathematical reasoning, multi-hop QA, and search-based QA, while achieving more than 10x efficiency over strong baselines. Together, MAS-Orchestra and MASBENCH enable better training and understanding of MAS in the pursuit of multi-agent intelligence.

We explore using a moderately sized large language model (GPT-J 6B parameters) to create a plan for a simulated robot to achieve 30 classes of goals in ScienceWorld, a text game simulator for elementary science experiments. Previously published empirical work claimed that large language models (LLMs) are a poor fit (Wang et al., 2022) compared to reinforcement learning. Using the Markov assumption (a single previous step), the LLM outperforms the reinforcement learning-based approach by a factor of 1.4. When we fill the LLM's input buffer with as many prior steps as possible, improvement rises to 3.5x. Even when training on only 6.5% of the training data, we observe a 2.2x improvement over the reinforcement-learning-based approach. Our experiments show that performance varies widely across the 30 classes of actions, indicating that averaging over tasks can hide significant performance issues. In work contemporaneous with ours, Lin et al. (2023) demonstrated a two-part approach (SwiftSage) that uses a small LLM (T5-large) complemented by OpenAI's massive LLMs to achieve outstanding results in ScienceWorld. Our 6-B parameter, single-stage GPT-J matches the performance of SwiftSage's two-stage architecture when it incorporates GPT-3.5 turbo which has 29-times more parameters than GPT-J.

Our goal is to deliver a new task and leaderboard to stimulate research on question answering and pre-trained language models (PTLMs) to understand a significant instructional document, e.g., an introductory college textbook or a manual. PTLMs have shown great success in many question-answering tasks, given significant supervised training, but much less so in zero-shot settings. We propose a new task that includes two college-level introductory texts in the social sciences (American Government 2e) and humanities (U.S. History), hundreds of true/false statements based on review questions written by the textbook authors, validation/development tests based on the first eight chapters of the textbooks, blind tests based on the remaining textbook chapters, and baseline results given state-of-the-art PTLMs. Since the questions are balanced, random performance should be ~50%. T5, fine-tuned with BoolQ achieves the same performance, suggesting that the textbook's content is not pre-represented in the PTLM. Taking the exam closed book, but having read the textbook (i.e., adding the textbook to T5's pre-training), yields at best minor improvement (56%), suggesting that the PTLM may not have "understood" the textbook (or perhaps misunderstood the questions). Performance is better (~60%) when the exam is taken open-book (i.e., allowing the machine to automatically retrieve a paragraph and use it to answer the question).

We describe Machine-Aided Script Curator (MASC), a system for human-machine collaborative script authoring. Scripts produced with MASC include (1) English descriptions of sub-events that comprise a larger, complex event; (2) event types for each of those events; (3) a record of entities expected to participate in multiple sub-events; and (4) temporal sequencing between the sub-events. MASC automates portions of the script creation process with suggestions for event types, links to Wikidata, and sub-events that may have been forgotten. We illustrate how these automations are useful to the script writer with a few case-study scripts.