Exploration-Exploitation Reshaping towards Efficient Reasoning for Large Language Models (2026.findings-acl)
Copied to clipboard
Yufeng Shi, Weilin Luo, Yuxiang Zhang, Zongmeng Zhang, Haoyang Liu, Yubing Wang, Bin Wang, Wengang Zhou, Houqiang Li
| Challenge: | Large Reasoning Models (LRMs) are constrained by the overthinking issue. |
| Approach: | They propose a policy optimization framework that reshapes the exploration and exploitation through two core components: self-imitation and self-guidance exploration. |
| Outcome: | The proposed model achieves superior reasoning efficiency without compromising overall accuracy. |
Similar Papers
Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches allocate an equal number of rollouts to all questions during the RL process, which is inefficient. |
| Approach: | They propose a mechanism for dynamically allocating rollout budgets based on the difficulty of the problems, enabling more efficient RL training. |
| Outcome: | The proposed model improves response precision while preserving exploratory ability to uncover potential correct pathways. |
LEASH: Adaptive Length Penalty and Reward Shaping for Efficient Large Reasoning Model (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to long reasoning traces are hard to tune and fail to adapt to evolving LLMs. |
| Approach: | They propose a reinforcement learning framework that optimizes the length of reasoning traces by a Lagrangian primal–dual method. |
| Outcome: | The proposed framework reduces the average reasoning length by 60% across diverse tasks while maintaining competitive performance. |
Correct, Concise and Complete: Multi-stage Training For Adaptive Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) increase test-time computation, often in the form of chain-of-thought (CoT) however, reasoning traces can become unnecessarily long, increasing computation costs without improving accuracy and sometimes even degrading performance. |
| Approach: | They propose a multi-stage efficient reasoning method that combines supervised fine-tuning with reinforcement learning using an adaptive length penalty. |
| Outcome: | The proposed method reduces response length by an average of 28% for 8B models and 40% for 32B models while incurring only minor performance drops of 1.6 and 2.5 points, respectively. |
MuTIS: Enhancing Reasoning Efficiency through Multi Turn Intervention Sampling in Reinforcement Learning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for training large reasoning models with long chain-of-thought (CoT) are limited by the number of parameters and the complexity of the model. |
| Approach: | They propose a framework that leverages multi-turn interventions to produce concise reasoning chains and demonstrates strong scalability. |
| Outcome: | The proposed framework breaks the accuracy-efficiency trade-off by producing concise reasoning chains and demonstrating strong scalability on 7B models. |
LEPO: Latent Reasoning Policy Optimization for Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing latent reasoning methods that use chain of thought (CoT) are limited to selecting one discrete token at each reasoning step, which potentially induces information loss. |
| Approach: | They propose a framework that injects controllable stochasticity into latent reasoning via Gumbel-Softmax, restoring LLMs' exploratory capacity and enhancing their compatibility with Reinforcement Learning (RL). |
| Outcome: | The proposed framework preserves richer information for more comprehensive reasoning and is compatible with Reinforcement Learning (RL). |
Unlocking Exploration in RLVR: Uncertainty-aware Advantage Shaping for Deeper Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Reinforcement Learning with Verifiable Rewards (RLVR) has shown significant promise for enhancing the reasoning capabilities of large language models (LLMs). |
| Approach: | They propose a model-free method that refines credit assignment by leveraging the model's internal uncertainty signals. |
| Outcome: | Extensive experiments on five mathematical reasoning benchmarks show that the proposed method outperforms strong RLVR baselines on multiple model scales, including 1.5B and 7B. |
SD-E2: Semantic Exploration for Reasoning Under Token Budgets (2026.findings-eacl)
Copied to clipboard
| Challenge: | Small language models struggle with complex reasoning because exploration is expensive under tight compute budgets. |
| Approach: | They propose a framework that makes exploration explicit by optimizing semantic diversity in generated reasoning trajectories. |
| Outcome: | The proposed framework surpasses Qwen2.5-3B-Instruct and strong GRPO baselines on GSM8K and improves on the harder AIME benchmark to 13.28% vs. base 6.74%. |
Reinforced Efficient Reasoning via Semantically Diverse Exploration (2026.acl-long)
Copied to clipboard
Ziqi Zhao, Zhaochun Ren, Jiahong Zou, Liu Yang, Zhiwei Xu, Xuri Ge, Zhumin Chen, Xinyu Ma, Daiting Shi, Shuaiqiang Wang, Dawei Yin, Xin Xin
| Challenge: | Existing methods for reinforcement learning with verifiable rewards suffer from limited exploration diversity and inefficient reasoning. |
| Approach: | They propose a method that rewards concise and correct reasoning while penalizing unnecessarily long reasoning chains. |
| Outcome: | Extensive experiments on Qwen and Llama models validate the effectiveness and efficiency of ROSE. |
Beyond High-Entropy Exploration: Correctness-Aware Low-Entropy Segment-Based Advantage Shaping for Reasoning LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent work studies RLVR through token entropy, arguing that high-entropies drive exploration and should receive stronger updates. |
| Approach: | They propose a correctness-aware reinforcement framework that performs fine-grained advantage modulation over low-entropy segments. |
| Outcome: | The proposed framework improves accuracy over strong RL baselines across three backbones and six math benchmarks while maintaining high-entropy exploration. |
EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning for LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to reinforcement learning with verifiable reward (RLVR) are limited by difficulty or lack of exploration. |
| Approach: | They propose a self-evolving curriculum learning framework based on chain-of-thought reasoning optimization that constrains exploration space by self-generating and verifying CoT trajectories. |
| Outcome: | The proposed framework enables LLMs to solve previously unsolved problems without external supervision and is compatible with various RL fine-tuning methods. |