A new framework called DecisionQE has been developed to measure the persuasive and compliant tendencies of large language models (LLMs) in group decision-making scenarios. Using the Werewolf game as a testbed, researchers found that while persuasive tendencies did not significantly improve group outcomes, compliant models showed more stable advantages in cooperation. The study suggests that LLMs can offer insights into sociological observations of language-mediated interactions and highlights the importance of incorporating behavioral tendencies into LLM safety evaluations. AI
IMPACT This research could lead to more robust safety evaluations for LLMs by quantifying their behavioral tendencies in social interactions.
RANK_REASON The cluster contains a research paper detailing a new framework for evaluating LLM behavior. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- DecisionQE
- Gotit.pub
- Humans
- Hugging Face
- Language Models
- ScienceCast
- Werewolf game
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
