Prolonged Reasoning Is Not All You Need: Certainty-Based Adaptive Routing for Efficient LLM/MLLM Reasoning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916748301697024 |
|---|---|
| author | Lu, Jinghui Yu, Haiyang Xu, Siliang Ran, Shiwei Tang, Guozhi Wang, Siqi Shan, Bin Fu, Teng Feng, Hao Tang, Jingqun Wang, Han Huang, Can |
| author_facet | Lu, Jinghui Yu, Haiyang Xu, Siliang Ran, Shiwei Tang, Guozhi Wang, Siqi Shan, Bin Fu, Teng Feng, Hao Tang, Jingqun Wang, Han Huang, Can |
| contents | Recent advancements in reasoning have significantly enhanced the capabilities of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) across diverse tasks. However, excessive reliance on chain-of-thought (CoT) reasoning can impair model performance and brings unnecessarily lengthened outputs, reducing efficiency. Our work reveals that prolonged reasoning does not universally improve accuracy and even degrade performance on simpler tasks. To address this, we propose Certainty-based Adaptive Reasoning (CAR), a novel framework that dynamically switches between short answers and long-form reasoning based on the model perplexity. CAR first generates a short answer and evaluates its perplexity, triggering reasoning only when the model exhibits low confidence (i.e., high perplexity). Experiments across diverse multimodal VQA/KIE benchmarks and text reasoning datasets show that CAR outperforms both short-answer and long-form reasoning approaches, striking an optimal balance between accuracy and efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_15154 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Prolonged Reasoning Is Not All You Need: Certainty-Based Adaptive Routing for Efficient LLM/MLLM Reasoning Lu, Jinghui Yu, Haiyang Xu, Siliang Ran, Shiwei Tang, Guozhi Wang, Siqi Shan, Bin Fu, Teng Feng, Hao Tang, Jingqun Wang, Han Huang, Can Computation and Language Artificial Intelligence Multimedia Recent advancements in reasoning have significantly enhanced the capabilities of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) across diverse tasks. However, excessive reliance on chain-of-thought (CoT) reasoning can impair model performance and brings unnecessarily lengthened outputs, reducing efficiency. Our work reveals that prolonged reasoning does not universally improve accuracy and even degrade performance on simpler tasks. To address this, we propose Certainty-based Adaptive Reasoning (CAR), a novel framework that dynamically switches between short answers and long-form reasoning based on the model perplexity. CAR first generates a short answer and evaluates its perplexity, triggering reasoning only when the model exhibits low confidence (i.e., high perplexity). Experiments across diverse multimodal VQA/KIE benchmarks and text reasoning datasets show that CAR outperforms both short-answer and long-form reasoning approaches, striking an optimal balance between accuracy and efficiency. |
| title | Prolonged Reasoning Is Not All You Need: Certainty-Based Adaptive Routing for Efficient LLM/MLLM Reasoning |
| topic | Computation and Language Artificial Intelligence Multimedia |
| url | https://arxiv.org/abs/2505.15154 |