HERO'S JOURNEY: Testing Complex Rule Induction with Text Games
- Published
- Source
- arXiv
- Paper number
- 299
- Field
- LLMs / NLP
- arXiv ID
- 2606.02556
Key points
- The paper introduces HERO'S JOURNEY, a benchmark for rule induction in goal-directed episodic tasks, where agents must infer hidden rules from demonstrations and act accordingly through multi-step execution.
- Evaluation of state-of-the-art LLMs found that the models show evidence of rule induction, but the ability is limited and uneven across tasks.
- Induction-specific steering methods improve performance on attribute tasks but do not reliably improve procedural tasks, suggesting that the gap in procedural induction remains an open problem.
Paper links
External research summaries. These are not HDATF publications or measured product results.