Post-training — page 1

• BDH-CQ: In-Context Learning with Recurrent Latent Reasoning
  arxiv:2608.09888 · 2026-08-10 · text available
• Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
  arxiv:2506.01939 · 2025-06-02 · text available
• Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
  arxiv:2509.14234 · 2025-09-17 · text available
• Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback
  arxiv:2506.03106 · 2025-06-03 · text available
• Curriculum Reinforcement Learning from Easy to Hard Tasks Improves LLM Reasoning
  arxiv:2506.06632 · 2025-06-07 · text available
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
  arxiv:2503.14476 · 2025-03-18 · text available
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
  arxiv:2402.03300 · 2024-02-05 · text available
• Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
  arxiv:2504.13837 · 2025-04-18 · text available
• Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning
  arxiv:2509.03646 · 2025-09-03 · text available
• ESPO: Entropy Importance Sampling Policy Optimization
  arxiv:2512.00499 · 2025-11-29 · text available
• Evolutionary Strategies lead to Catastrophic Forgetting in LLMs
  arxiv:2601.20861 · 2026-01-28 · text available
• Fine-Tuning Fixes Mode Collapse and Over-Dispersion in LLMs
  arxiv:2609.16454 · 2026-09-15 · text available
• Finetuning with Sampling: SFT Learns Better Than You Think
  arxiv:2610.02140 · 2026-10-01 · text available
• From $f(x)$ and $g(x)$ to $f(g(x))$: LLMs Learn New Skills in RL by Composing Old Ones
  arxiv:2509.25123 · 2025-09-29 · text available
• From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning
  arxiv:2605.22074 · 2026-05-21 · text available
• GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
  arxiv:2507.19457 · 2025-07-25 · text available
• Gradient Regularization Mitigates Reward Hacking in Reinforcement Learning from Human Feedback and Verifiable Rewards
  arxiv:2602.18037 · 2026-02-20 · text available
• Group Sequence Policy Optimization
  arxiv:2507.18071 · 2025-07-24 · text available
• Group-in-Group Policy Optimization for LLM Agent Training
  arxiv:2505.10978 · 2025-05-16 · text available
• GRPO++: Tricks for Making RL Actually Work (Cameron Wolfe)
  url:cameronrwolfe.substack.com/p/grpo-tricks · 2026-01-05 · metadata only
• GRPO-VPS: Enhancing Group Relative Policy Optimization with Verifiable Process Supervision for Effective Reasoning
  arxiv:2604.20659 · 2026-04-22 · text available
• iGRPO: Self-Feedback-Driven LLM Reasoning
  arxiv:2602.09000 · 2026-02-09 · text available
• It Takes Two: Your GRPO Is Secretly DPO
  arxiv:2510.00977 · 2025-10-01 · text available
• Klear-Reasoner: Advancing Reasoning Capability via Gradient-Preserving Clipping Policy Optimization
  arxiv:2508.07629 · 2025-08-11 · text available