Not All Reasoning Generalizes: A Case Study on Learning Generalizable Reasoning Strategies in RLVR
cs.LG
Submitted: 2026-11-02
Updated: 2026-11-02
Code: https://github.com/ziqipang/rl_meta_
Terminology
Sources
- The Llama 3 Herd of Models
- Kimi k1.5: Scaling Reinforcement Learning with LLMs
- TinyGSM: achieving >80% on GSM8k with small language models
- Length-Unbiased Sequence Policy Optimization: Revealing and Controlling Response Length Variation in RLVR
- In-context Learning and Induction Heads
- Proximal Policy Optimization Algorithms
- Spurious Rewards: Rethinking Training Signals in RLVR
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization
- RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning
- RAGEN-2: Reasoning Collapse in Agentic RL
- Measuring short-form factuality in large language models
- The Invisible Leash: Why RLVR May or May Not Escape Its Origin
- Qwen3 Technical Report
- Learning Composable Chains-of-Thought
- On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
- Improving Sampling Efficiency in RLVR through Adaptive Rollout and Response Reuse
- Group Sequence Policy Optimization
- GSM-Infinite: How Do Your LLMs Behave over Infinitely Increasing Context Length and Reasoning Complexity?
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