13 papers
Stabilizing Efficient Reasoning with Step-Level Advantage Selection
Han Wang, Xiaodong Yu, Jialian Wu +4
Large language models (LLMs) achieve strong reasoning performance by allocating substantial computation at inference time, often generating long and verbose reasoning traces. While…
DeepCompress: A Dual Reward Strategy for Dynamically Exploring and Compressing Reasoning Chains
Tian Liang, Wenxiang Jiao, Zhiwei He +3
Large Reasoning Models (LRMs) have demonstrated impressive capabilities but suffer from cognitive inefficiencies like "overthinking" simple problems and "underthinking" complex one…
Draft Model Knows When to Stop: Self-Verification Speculative Decoding for Long-Form Generation
Ziyin Zhang, Jiahao Xu, Tian Liang +4
Conventional speculative decoding (SD) methods utilize a predefined length policy for proposing drafts, which implies the premise that the target model smoothly accepts the propose…
DeepTheorem: Advancing LLM Reasoning for Theorem Proving Through Natural Language and Reinforcement Learning
Ziyin Zhang, Jiahao Xu, Zhiwei He +10
Theorem proving serves as a major testbed for evaluating complex reasoning abilities in large language models (LLMs). However, traditional automated theorem proving (ATP) approache…
Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training
Mengru Wang, Xingyu Chen, Yue Wang +12
Mixture-of-Experts (MoE) architectures within Large Reasoning Models (LRMs) have achieved impressive reasoning capabilities by selectively activating experts to facilitate structur…
DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning
Zhiwei He, Tian Liang, Jiahao Xu +12
Reinforcement learning (RL) with large language models shows promise in complex reasoning. However, its progress is hindered by the lack of large-scale training data that is suffic…