Negative-Only Policy Optimization for One-Sided Verifiable Rewards
cs.LG, cs.CL
Submitted: 2026-11-12
Updated: 2026-11-12
Terminology
Sources
- The Unreasonable Effectiveness of Entropy Minimization in LLM Reasoning
- Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers
- Evaluating Large Language Models Trained on Code
- Training Verifiers to Solve Math Word Problems
- Noise-corrected GRPO: From Noisy Rewards to Unbiased Gradients
- Who's Harry Potter? Approximate Unlearning in LLMs
- Don't Waste Mistakes: Leveraging Negative RL-Groups via Confidence Reweighting
- R-Zero: Self-Evolving Reasoning LLM from Zero Data
- Kimi k1.5: Scaling Reinforcement Learning with LLMs
- Tulu 3: Pushing Frontiers in Open Language Model Post-Training
- TACO: Topics in Algorithmic COde generation dataset
- Tool Verification for Test-Time Reinforcement Learning
- NGRPO: Negative-enhanced Group Relative Policy Optimization
- OpenAI o1 System Card
- Rate or Fate? RLV$^\varepsilon$R: Reinforcement Learning with Verifiable Noisy Rewards
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- LADDER: Self-Improving LLMs Through Recursive Problem Decomposition
- Rethinking Sample Polarity in Reinforcement Learning with Verifiable Rewards
- Qwen3 Technical Report
- Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis