Cross-Centaur: Modeling Individual Decision-Making Across Psychological Tasks with Language Models

Key points

• Large language models (LLMs) are increasingly used as computational models of human behavior.1
• We train Cross-Centaur to predict a participant's target-task actions from a target-task description and the source-task description paired with the same participant's source-task behavior.1
• Probing analyses suggest that its hidden states encode participant-level behavioral tendencies that transfer across tasks.1
• These results highlight the potential of cross-task individual-level data for predicting human decision-making with language models.1

Sources (1)

• [1]Cross-Centaur: Modeling Individual Decision-Making Across Psychological Tasks with Language Models
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 08:55 AM

  We introduce Cross-Psych-101, a dataset of 591 participants and 261{,}551 valid decisions across 12 Bandit, Tradeoff, Structure, and Social tasks.

  Large language models (LLMs) are increasingly used as computational models of human behavior.

Extractive summary: sentences quoted from the sources.