LLMs improve Theory-of-Mind reasoning with explicit if-then rules · 1 source tracked

Researchers have developed a method to improve Theory-of-Mind (ToM) reasoning in large language models by generating explicit if-then rules that connect perception, belief, and action. Using Group Relative Policy Optimization (GRPO), they trained models like Qwen2.5 and Llama 3 on ToM benchmarks, finding that this rule-based approach surpasses other methods such as prompting and supervised fine-tuning. Furthermore, rules generated by larger models demonstrated transferable knowledge, enhancing the performance of smaller models like GPT-4o mini. AI

IMPACT This research could lead to LLMs that better understand and predict human behavior, improving applications in areas like human-robot interaction and personalized AI assistants.

RANK_REASON The cluster contains an academic paper detailing a new method for improving LLM reasoning capabilities. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• arXiv
• GPT-4o mini
• Group Relative Policy Optimization
• Grpo
• Llama 3
• Qwen2.5
• theory of mind

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →