Streaming Hallucination Detection in Long Chain-of-Thought Reasoning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908748347342848 |
|---|---|
| author | Lu, Haolang Pan, Minghui Li, Ripeng Nan, Guoshun Zhuang, Jialin Zhao, Zijie Sun, Zhongxiang Wang, Kun Liu, Yang |
| author_facet | Lu, Haolang Pan, Minghui Li, Ripeng Nan, Guoshun Zhuang, Jialin Zhao, Zijie Sun, Zhongxiang Wang, Kun Liu, Yang |
| contents | Long chain-of-thought (CoT) reasoning improves the performance of large language models, yet hallucinations in such settings often emerge subtly and propagate across reasoning steps. We suggest that hallucination in long CoT reasoning is better understood as an evolving latent state rather than a one-off erroneous event. Accordingly, we treat step-level hallucination judgments as local observations and introduce a cumulative prefix-level hallucination signal that tracks the global evolution of the reasoning state over the entire trajectory. Overall, our approach enables streaming hallucination detection in long CoT reasoning, providing real-time, interpretable evidence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_02170 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Streaming Hallucination Detection in Long Chain-of-Thought Reasoning Lu, Haolang Pan, Minghui Li, Ripeng Nan, Guoshun Zhuang, Jialin Zhao, Zijie Sun, Zhongxiang Wang, Kun Liu, Yang Artificial Intelligence Long chain-of-thought (CoT) reasoning improves the performance of large language models, yet hallucinations in such settings often emerge subtly and propagate across reasoning steps. We suggest that hallucination in long CoT reasoning is better understood as an evolving latent state rather than a one-off erroneous event. Accordingly, we treat step-level hallucination judgments as local observations and introduce a cumulative prefix-level hallucination signal that tracks the global evolution of the reasoning state over the entire trajectory. Overall, our approach enables streaming hallucination detection in long CoT reasoning, providing real-time, interpretable evidence. |
| title | Streaming Hallucination Detection in Long Chain-of-Thought Reasoning |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2601.02170 |