Entropy-based Exploration Conduction for Multi-step Reasoning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915348606877696 |
|---|---|
| author | Zhang, Jinghan Wang, Xiting Mo, Fengran Zhou, Yeyang Gao, Wanfu Liu, Kunpeng |
| author_facet | Zhang, Jinghan Wang, Xiting Mo, Fengran Zhou, Yeyang Gao, Wanfu Liu, Kunpeng |
| contents | Multi-step processes via large language models (LLMs) have proven effective for solving complex reasoning tasks. However, the depth of exploration of the reasoning procedure can significantly affect the task performance. Existing methods to automatically decide the depth often lead to high cost and a lack of flexibility. To address these issues, we propose Entropy-based Exploration Depth Conduction (Entro-duction), a novel method that dynamically adjusts the exploration depth during multi-step reasoning by monitoring LLM's output entropy and variance entropy. We employ these two features to capture the model's uncertainty of the current step and the fluctuation of uncertainty across consecutive reasoning steps. Based on the observed entropy changes, the LLM selects whether to deepen, expand, or stop exploration according to the probability, which facilitates the trade-off between the reasoning accuracy and exploration effectiveness. Experimental results across four benchmark datasets demonstrate the efficacy of Entro-duction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_15848 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Entropy-based Exploration Conduction for Multi-step Reasoning Zhang, Jinghan Wang, Xiting Mo, Fengran Zhou, Yeyang Gao, Wanfu Liu, Kunpeng Artificial Intelligence Computation and Language Multi-step processes via large language models (LLMs) have proven effective for solving complex reasoning tasks. However, the depth of exploration of the reasoning procedure can significantly affect the task performance. Existing methods to automatically decide the depth often lead to high cost and a lack of flexibility. To address these issues, we propose Entropy-based Exploration Depth Conduction (Entro-duction), a novel method that dynamically adjusts the exploration depth during multi-step reasoning by monitoring LLM's output entropy and variance entropy. We employ these two features to capture the model's uncertainty of the current step and the fluctuation of uncertainty across consecutive reasoning steps. Based on the observed entropy changes, the LLM selects whether to deepen, expand, or stop exploration according to the probability, which facilitates the trade-off between the reasoning accuracy and exploration effectiveness. Experimental results across four benchmark datasets demonstrate the efficacy of Entro-duction. |
| title | Entropy-based Exploration Conduction for Multi-step Reasoning |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2503.15848 |