Not All Thoughts are Generated Equal: Efficient LLM Reasoning via Multi-Turn Reinforcement Learning

• URL: http://arxiv.org/abs/2505.11827v1
• Date: Sat, 17 May 2025 04:26:39 GMT
• Title: Not All Thoughts are Generated Equal: Efficient LLM Reasoning via Multi-Turn Reinforcement Learning
• Authors: Yansong Ning, Wei Li, Jun Fang, Naiqiang Tan, Hao Liu,
• Abstract summary: Long chain-of-thought (CoT) from large language models (LLMs) is an emerging strategy to improve the reasoning efficiency of LLMs.<n>Existing studies equally compress all thoughts within a long CoT, hindering more concise and effective reasoning.<n>We propose Long$otimes$Short, an efficient reasoning framework that enables two LLMs to collaboratively solve the problem.
• Score: 12.830215971176806
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Compressing long chain-of-thought (CoT) from large language models (LLMs) is an emerging strategy to improve the reasoning efficiency of LLMs. Despite its promising benefits, existing studies equally compress all thoughts within a long CoT, hindering more concise and effective reasoning. To this end, we first investigate the importance of different thoughts by examining their effectiveness and efficiency in contributing to reasoning through automatic long CoT chunking and Monte Carlo rollouts. Building upon the insights, we propose a theoretically bounded metric to jointly measure the effectiveness and efficiency of different thoughts. We then propose Long$\otimes$Short, an efficient reasoning framework that enables two LLMs to collaboratively solve the problem: a long-thought LLM for more effectively generating important thoughts, while a short-thought LLM for efficiently generating remaining thoughts. Specifically, we begin by synthesizing a small amount of cold-start data to fine-tune LLMs for long-thought and short-thought reasoning styles, respectively. Furthermore, we propose a synergizing-oriented multi-turn reinforcement learning, focusing on the model self-evolution and collaboration between long-thought and short-thought LLMs. Experimental results show that our method enables Qwen2.5-7B and Llama3.1-8B to achieve comparable performance compared to DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B, while reducing token length by over 80% across the MATH500, AIME24/25, AMC23, and GPQA Diamond benchmarks. Our data and code are available at https://github.com/yasNing/Long-otimes-Short/.

Related papers

• Learn to Reason Efficiently with Adaptive Length-based Reward Shaping [23.626013831589212]
  Large Reasoning Models (LRMs) have shown remarkable capabilities in solving complex problems through reinforcement learning (RL)<n>We present a unified framework that formulates various efficient reasoning methods through the lens of length-based reward shaping.<n>Experiments on DeepSeek-R1-Distill-Qwen-1.5B, DeepSeek-R1-Distill-Qwen-7B, and DeepSeek-R1-Distill-Qwen-32B show that our approach significantly enhances both reasoning performance and response length efficiency.
  arXiv  Detail & Related papers  (2025-05-21T15:03:26Z)
• Thinkless: LLM Learns When to Think [57.857534644932194]
  Reasoning Language Models, capable of extended chain-of-thought reasoning, have demonstrated remarkable performance on tasks requiring complex logical inference.<n>We propose Thinkless, a learnable framework that empowers an LLM to adaptively select between short-form and long-form reasoning.<n>On several benchmarks such as Minerva Algebra, MATH-500, and GSM8K, Thinkless is able to reduce the usage of long-chain thinking by 50% - 90%.
  arXiv  Detail & Related papers  (2025-05-19T17:24:16Z)
• Thinking Short and Right Over Thinking Long: Serving LLM Reasoning Efficiently and Accurately [29.018731931275138]
  Large Language Models (LLMs) can gain better capabilities by generating Chain-of-Thought reasoning to respond a given request.<n>However, when incorporating the two scaling dimensions, the system efficiency is dampened significantly for two reasons.<n>We present SART, a serving framework for efficient and accurate LLM reasoning.
  arXiv  Detail & Related papers  (2025-05-19T16:34:56Z)
• Learning to Think: Information-Theoretic Reinforcement Fine-Tuning for LLMs [25.03191529055168]
  Large language models (LLMs) excel at complex tasks thanks to advances in reasoning abilities.<n>Existing methods overlook the trade-off between reasoning effectiveness and computational efficiency.<n>We propose Learning to Think to make the models achieve optimal reasoning with fewer tokens.
  arXiv  Detail & Related papers  (2025-05-15T15:40:25Z)
• ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning [68.02825465552779]
  We present ThinkPrune, a simple yet effective method for pruning the thinking length for long-thinking LLMs.<n>We show that ThinkPrune results in a remarkable performance-length tradeoff -- on the AIME24 dataset, the reasoning length of DeepSeek-R1-Distill-Qwen-1.5B can be reduced by half with only 2% drop in performance.
  arXiv  Detail & Related papers  (2025-04-02T01:59:26Z)
• Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration [70.09561665520043]
  We propose a novel framework for multi-agent collaboration that introduces Reinforced Advantage feedback (ReAd) for efficient self-refinement of plans. We provide theoretical analysis by extending advantage-weighted regression in reinforcement learning to multi-agent systems. Experiments on Over-AI and a difficult variant of RoCoBench show that ReAd surpasses baselines in success rate, and also significantly decreases the interaction steps of agents.
  arXiv  Detail & Related papers  (2024-05-23T08:33:19Z)
• Generating Chain-of-Thoughts with a Pairwise-Comparison Approach to Searching for the Most Promising Intermediate Thought [70.30423016640749]
  Chain-of-thoughts (CoT) methods were proposed to guide large language models to reason step-by-step, enabling problem solving from simple to complex. The evaluation from the large language model (LLMs) is typically noisy and unreliable, potentially misleading the generation process in selecting promising intermediate thoughts. In this paper, motivated by Vapnik's principle, we use pairwise-comparison evaluation instead of point-wise scoring to search for promising intermediate thoughts.
  arXiv  Detail & Related papers  (2024-02-10T09:51:03Z)
• Efficient Tool Use with Chain-of-Abstraction Reasoning [63.08202389132155]
  Large language models (LLMs) need to ground their reasoning to real-world knowledge.<n>There remains challenges for fine-tuning LLM agents to invoke tools in multi-step reasoning problems.<n>We propose a new method for LLMs to better leverage tools in multi-step reasoning.
  arXiv  Detail & Related papers  (2024-01-30T21:53:30Z)
• Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselves [57.974103113675795]
  We present a method named Rephrase and Respond' (RaR) which allows Large Language Models to rephrase and expand questions posed by humans. RaR serves as a simple yet effective prompting method for improving performance. We show that RaR is complementary to the popular Chain-of-Thought (CoT) methods, both theoretically and empirically.
  arXiv  Detail & Related papers  (2023-11-07T18:43:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.