Credit Without Ground Truth: Auditing Step-Level Credit Assignment in LLM Agents Against Executed Replay
无真实值的信用:针对执行重放的大语言模型智能体的步骤级信用分配审计
机构 * University of Southern California(南加州大学)
AI总结 该研究针对LLM智能体,在ALFWorld环境中审计步骤级信用分配,发现现有信用信号无法识别关键步骤,提出需匹配有效样本量比较信用规则。
Comments 49 pages, 7 figures. Pre-registered; frozen analysis plans and prompts included in the appendices. Under review