Explore Briefly, Then Decide: Mitigating LLM Overthinking via Cumulative Entropy Regulation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910062224605184 |
|---|---|
| author | Bin, Yi Jiang, Tianyi Ding, Yujuan Zhu, Kainian Ma, Fei Song, Jingkuan Yang, Yang Shen, Heng Tao |
| author_facet | Bin, Yi Jiang, Tianyi Ding, Yujuan Zhu, Kainian Ma, Fei Song, Jingkuan Yang, Yang Shen, Heng Tao |
| contents | Large Language Models (LLMs) have demonstrated remarkable reasoning abilities on complex problems using long Chain-of-Thought (CoT) reasoning. However, they often suffer from overthinking, meaning generating unnecessarily lengthy reasoning steps for simpler problems. This issue may degrade the efficiency of the models and make them difficult to adapt the reasoning depth to the complexity of problems. To address this, we introduce a novel metric Token Entropy Cumulative Average (TECA), which measures the extent of exploration throughout the reasoning process. We further propose a novel reasoning paradigm named "Explore Briefly, Then Decide", with an associated Cumulative Entropy Regulation (CER) mechanism. This paradigm leverages TECA to help the model dynamically determine the optimal point to conclude its thought process and provide a final answer, thus achieving efficient reasoning. Experimental results across diverse mathematical benchmarks show that our approach substantially mitigates overthinking without sacrificing problem solving ability. With our thinking paradigm, the average response length decreases by up to 71% on simpler datasets, demonstrating the effectiveness of our method in creating a more efficient and adaptive reasoning process. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_02249 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Explore Briefly, Then Decide: Mitigating LLM Overthinking via Cumulative Entropy Regulation Bin, Yi Jiang, Tianyi Ding, Yujuan Zhu, Kainian Ma, Fei Song, Jingkuan Yang, Yang Shen, Heng Tao Computation and Language Artificial Intelligence Machine Learning Large Language Models (LLMs) have demonstrated remarkable reasoning abilities on complex problems using long Chain-of-Thought (CoT) reasoning. However, they often suffer from overthinking, meaning generating unnecessarily lengthy reasoning steps for simpler problems. This issue may degrade the efficiency of the models and make them difficult to adapt the reasoning depth to the complexity of problems. To address this, we introduce a novel metric Token Entropy Cumulative Average (TECA), which measures the extent of exploration throughout the reasoning process. We further propose a novel reasoning paradigm named "Explore Briefly, Then Decide", with an associated Cumulative Entropy Regulation (CER) mechanism. This paradigm leverages TECA to help the model dynamically determine the optimal point to conclude its thought process and provide a final answer, thus achieving efficient reasoning. Experimental results across diverse mathematical benchmarks show that our approach substantially mitigates overthinking without sacrificing problem solving ability. With our thinking paradigm, the average response length decreases by up to 71% on simpler datasets, demonstrating the effectiveness of our method in creating a more efficient and adaptive reasoning process. |
| title | Explore Briefly, Then Decide: Mitigating LLM Overthinking via Cumulative Entropy Regulation |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2510.02249 |