paper-with-me

홈 › Papers

Optimizing Reasoning Efficiency through Prompt Difficulty Prediction

2025-11-05 · Bo Zhao, Berkcan Kapusuzoglu, Kartik Balasubramaniam, Sambit Sahu, Supriyo Chakraborty, Genta Indra Winata arxiv

Reasoning language models perform well on complex tasks but are costly to deploy due to their size and long reasoning traces. We propose a routing approach that assigns each problem to the smallest model likely to solve it, reducing compute without sacrificing accuracy. Using intermediate representations from s1.1-32B, we train lightweight predictors of problem difficulty or model correctness to guide routing across a pool of reasoning models. On diverse math benchmarks, routing improves efficiency over random assignment and matches s1.1-32B's performance while using significantly less compute. Our results demonstrate that difficulty-aware routing is effective for cost-efficient deployment of reasoning models.

📄 PDF Abstract BibTeX arXiv:2511.03808

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Kalman Meets Curriculum: Efficient Dynamic Prompt Selection for Adaptive RL Finetuning

2026-07-30 · Haodong Zhu, Yangyang Ren, Yanjing Li, Sheng Xu 외 arxiv

Reinforcement learning (RL) finetuning significantly enhances the reasoning capabilities of large language models (LLMs), yet its effectiveness critically depends on selecting prompts of appropriate difficulty for the cu…

Reinforcement Learning

Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Models

2026-02-02 · Yun Qu, Qi Wang, Yixiu Mao, Heming Zou 외 arxiv

Reinforcement learning enhances the reasoning capabilities of large language models but often involves high computational costs due to rollout-intensive optimization. Online prompt selection presents a plausible solution…

Reinforcement LearningBayesian Inference

AdaptThink: Reasoning Models Can Learn When to Think

2025-05-19 · Jiajie Zhang, Nianyi Lin, Lei Hou, Ling Feng 외

Recently, large reasoning models have achieved impressive performance on various tasks by employing human-like deep thinking. However, the lengthy thinking process substantially increases inference overhead, making effic…

Math

SPEED-RL: Faster Training of Reasoning Models via Online Curriculum Learning

2025-06-10 · Ruiqi Zhang, Daman Arora, Song Mei, Andrea Zanette

Training large language models with reinforcement learning (RL) against verifiable rewards significantly enhances their reasoning abilities, yet remains computationally expensive due to inefficient uniform prompt samplin…

Reinforcement Learning (RL)

PATS: Process-Level Adaptive Thinking Mode Switching

2025-05-25 · Yi Wang, Junxiao Liu, Shimao Zhang, Jiajun Chen 외

Current large-language models (LLMs) typically adopt a fixed reasoning strategy, either simple or complex, for all questions, regardless of their difficulty. This neglect of variation in task and reasoning process comple…

Computational Efficiency