| title | When Should Models Change Their Minds? Contextual Belief Management in Large Language Models |
|---|---|
| authors | Haoming Xu, Weihong Xu, Zongrui Li, Mengru Wang, Yunzhi Yao, Chiyu Wu, Jin Shang, Yu Gong, Shumin Deng |
| date | 2026-05-28 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.30219 |
| affiliation | Zhejiang University |
| upvotes | 22 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Long-horizon interactions require language models to manage accumulating information: when to update their state, when to preserve their state, and what to ignore. We study this challenge as Contextual Belief Management (CBM): maintaining a predicted belief state aligned with formal evidence while isolating task-irrelevant noise. To make CBM measurable, we introduce BeliefTrack, a closed-world benchmark spanning Rule Discovery and Circuit Diagnosis, where a finite belief space and symbolic verifiers enable exact turn-level evaluation. BeliefTrack diagnoses three failures: Failed Stay, Failed Update, and Failed Isolation. Across multiple LLMs, vanilla models exhibit severe CBM failures, while explicit belief-tracking prompts provide limited gains. In contrast, reinforcement learning with belief-state rewards reduces failure rates by 70.9\% on average. Further probing reveals latent belief-state dynamics behind these failures, and representation-level steering reduces failure rates by 46.1\% across two tasks\footnote{Code is coming soon at https://github.com/zjunlp/CBM. |
cite
@misc{xu2026when,
title={When Should Models Change Their Minds? Contextual Belief Management in Large Language Models},
author={Haoming Xu and Weihong Xu and Zongrui Li and Mengru Wang and Yunzhi Yao and Chiyu Wu and Jin Shang and Yu Gong and Shumin Deng},
year={2026},
eprint={2605.30219},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.30219}
}