What Will Post-Training Fix? Per-Problem Gains Are Shared Across Independent RL Runs, and Existing Checkpoints Predict Them Better Than A Priori Signals
cs.LG
Submitted: 2026-10-04
Updated: 2026-10-04
Sources
- MathArena: Evaluating LLMs on Uncontaminated Math Competitions
- Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR
- The Unlearnability Phenomenon in RLVR for Language Models
- CircuitLens: Reasoning Circuits as Data Selection Signals for Reinforcement Learning with Verifiable Rewards
- The Model Knows Another Way: Strategy Switching for Effective RLVR Exploration
- Reinforcement Learning for Reasoning in Small LLMs: What Works and What Doesn't
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Rethinking On-Policy Distillation of Large Language Models II: One Training Example
- Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language Models
- JustRL: Scaling a 1.5B LLM with a Simple RL Recipe
- OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems
- Measuring Mathematical Problem Solving With the MATH Dataset
- Solving Without Stopping: On-Policy Distillation at Small Scale
- Beyond Imitation: Auditing the Recoverability of Reasoning in Distilled Models
- DataFlex-RL: An Evaluation Platform for RLVR Data Policies
- ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
- Understanding R1-Zero-Like Training: A Critical Perspective
- Learning What Reinforcement Learning Can't: Interleaved Online Fine-Tuning for Hardest Questions
- Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems
- Spurious Rewards: Rethinking Training Signals in RLVR
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks