Taming AI Bots: Controllability of Neural States in Large Language Models
Stefano Soatto, Paulo Tabuada, Pratik Chaudhari, Tian Yu Liu
专题命中
其他LLM
:large language model(title,abstract);language model(title,abstract);LLM(abstract);分类 cs.CL、cs.AI、cs.LG
CommentsTLDR: AI Bots are stochastic dynamical systems whose mental state can be controlled by both the user and the designer. The space of meanings, defined as equivalence classes of sentences, is learned during fine-tuning with human supervision, and safeguarding can be designed into the bot by establishing controls both at its input and output
The Only Way is Ethics: A Guide to Ethical Research with Large Language Models
Eddie L. Ungless, Nikolas Vitsakis, Zeerak Talat, James Garforth, Björn Ross, Arno Onken, Atoosa Kasirzadeh, Alexandra Birch
专题命中
其他LLM
:large language model(title,abstract);language model(title,abstract);LLM(abstract,comments);分类 cs.CL、cs.AI
CommentsAccepted to COLING '25. This paper is the condensed pocket guide to accompany our full LLM Ethics Whitepaper, available at arXiv:2410.19812, and at https://github.com/MxEddie/Ethics-Whitepaper for suggested revisions
Comments16 pages main paper, 19 pages supplementary material included as ancillary file. Major revision with substantially reorganized and streamlined presentation; expanded held-out prediction, population-composition, input-sensitivity, and composition analyses; includes an MBPP corruption audit and updated related work