Image de couverture de livre

The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhu, Xinyu, Xia, Mengzhou, Wei, Zhepei, Chen, Wei-Lin, Chen, Danqi, Meng, Yu
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

SimPO: Simple Preference Optimization with a Reference-Free Reward
par: Meng, Yu, et autres
Publié: (2024)

InstructRAG: Instructing Retrieval-Augmented Generation via Self-Synthesized Rationales
par: Wei, Zhepei, et autres
Publié: (2024)

You Only Need Minimal RLVR Training: Extrapolating LLMs via Rank-1 Trajectories
par: Wei, Zhepei, et autres
Publié: (2026)

Lory: Fully Differentiable Mixture-of-Experts for Autoregressive Language Model Pre-training
par: Zhong, Zexuan, et autres
Publié: (2024)

Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning
par: Xia, Mengzhou, et autres
Publié: (2023)

Do LLM Evaluators Prefer Themselves for a Reason?
par: Chen, Wei-Lin, et autres
Publié: (2025)

AdaDecode: Accelerating LLM Decoding with Adaptive Layer Parallelism
par: Wei, Zhepei, et autres
Publié: (2025)

LESS: Selecting Influential Data for Targeted Instruction Tuning
par: Xia, Mengzhou, et autres
Publié: (2024)

ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning
par: Lin, Zihan, et autres
Publié: (2026)

How to Train Long-Context Language Models (Effectively)
par: Gao, Tianyu, et autres
Publié: (2024)

LitSearch: A Retrieval Benchmark for Scientific Literature Search
par: Ajith, Anirudh, et autres
Publié: (2024)

Aligning Large Language Models via Fully Self-Synthetic Data
par: Yin, Shangjian, et autres
Publié: (2025)

Detecting Pretraining Data from Large Language Models
par: Shi, Weijia, et autres
Publié: (2023)

Rethinking Mixture-of-Agents: Is Mixing Different Large Language Models Beneficial?
par: Li, Wenzhe, et autres
Publié: (2025)

DRAGON: Guard LLM Unlearning in Context via Negative Detection and Reasoning
par: Wang, Yaxuan, et autres
Publié: (2025)

Evaluating Large Language Models at Evaluating Instruction Following
par: Zeng, Zhiyuan, et autres
Publié: (2023)

The Heuristic Core: Understanding Subnetwork Generalization in Pretrained Language Models
par: Bhaskar, Adithya, et autres
Publié: (2024)

How Off-Policy Can GRPO Be? Mu-GRPO for Efficient LLM Reinforcement Learning
par: Tian, Minghao, et autres
Publié: (2026)

Trainable Transformer in Transformer
par: Panigrahi, Abhishek, et autres
Publié: (2023)

Retaining by Doing: The Role of On-Policy Data in Mitigating Forgetting
par: Chen, Howard, et autres
Publié: (2025)

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
par: Wang, Shenzhi, et autres
Publié: (2025)

MoDoMoDo: Multi-Domain Data Mixtures for Multimodal LLM Reinforcement Learning
par: Liang, Yiqing, et autres
Publié: (2025)

CultureLLM: Incorporating Cultural Differences into Large Language Models
par: Li, Cheng, et autres
Publié: (2024)

Teaching LLM to Reason: Reinforcement Learning from Algorithmic Problems without Code
par: Bao, Keqin, et autres
Publié: (2025)

DAPO: An Open-Source LLM Reinforcement Learning System at Scale
par: Yu, Qiying, et autres
Publié: (2025)

Dynamic Thinking-Token Selection for Efficient Reasoning in Large Reasoning Models
par: Guo, Zhenyuan, et autres
Publié: (2026)

CLUE: Conflict-guided Localization for LLM Unlearning Framework
par: Chen, Hang, et autres
Publié: (2025)

WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning
par: Wei, Zhepei, et autres
Publié: (2025)

Extracting Rule-based Descriptions of Attention Features in Transformers
par: Friedman, Dan, et autres
Publié: (2025)

QuRating: Selecting High-Quality Data for Training Language Models
par: Wettig, Alexander, et autres
Publié: (2024)

Harnessing Negative Signals: Reinforcement Distillation from Teacher Data for LLM Reasoning
par: Xu, Shuyao, et autres
Publié: (2025)

Incentivizing Agentic Reasoning in LLM Judges via Tool-Integrated Reinforcement Learning
par: Xu, Ran, et autres
Publié: (2025)

TruthRL: Incentivizing Truthful LLMs via Reinforcement Learning
par: Wei, Zhepei, et autres
Publié: (2025)

NFT: Bridging Supervised Learning and Reinforcement Learning in Math Reasoning
par: Chen, Huayu, et autres
Publié: (2025)

Save the Good Prefix: Precise Error Penalization via Process-Supervised RL to Enhance LLM Reasoning
par: Liu, Haolin, et autres
Publié: (2026)

Representing Rule-based Chatbots with Transformers
par: Friedman, Dan, et autres
Publié: (2024)

ReMix: Reinforcement routing for mixtures of LoRAs in LLM finetuning
par: Qiu, Ruizhong, et autres
Publié: (2026)

On Designing Effective RL Reward at Training Time for LLM Reasoning
par: Gao, Jiaxuan, et autres
Publié: (2024)

C3oT: Generating Shorter Chain-of-Thought without Compromising Effectiveness
par: Kang, Yu, et autres
Publié: (2024)

Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning
par: Zhang, Ruiqi, et autres
Publié: (2024)