Collaborative Theory of Mind: Learning to Self-Revise by Anticipating User Responses

Key points

• Large language models (LLMs) demonstrate strong individual task performance, yet still struggle to collaborate with users over extended, multi-turn interactions, for example by attempting solutions before understanding the user's needs or by failing to revise mistaken assumptions after receiving feedback.1
• We combine this prediction objective with group-relative policy optimization (GRPO) to optimize both task performance and interaction efficiency.1
• Across mathematical reasoning, coding, and writing tasks, CToM achieves higher collaborative task performance than all baselines; for example, on a collaborative mathematical reasoning task, it improves final accuracy from 74.1% to 79.6% compared with naive GRPO.1
• Behavioral analyses of interaction trajectories show that user-response predictions become more accurate after training, and that more accurate predictions are associated with larger improvements in collaborative task performance through self-revision.1

Sources (1)

• [1]Collaborative Theory of Mind: Learning to Self-Revise by Anticipating User Responses
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 02:01 AM

  We propose Collaborative Theory of Mind (CToM), a framework that addresses these challenges by explicitly training LLMs to anticipate user responses and to use those predictions to improve their own behavior.

  Large language models (LLMs) demonstrate strong individual task performance, yet still struggle to collaborate with users over extended, multi-turn interactions, for example by attempting solutions before understanding the user's needs or by failing to revise mistaken assumptions after receiving feedback.

Extractive summary: sentences quoted from the sources.