Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR
cs.AI
Submitted: 2026-09-30
Updated: 2026-09-30
Sources
- Off-Context GRPO: Learning to Reason on Hard Problems using Privileged Information
- Prompt replay: speeding up grpo with on-policy reuse of high-signal prompts
- Max Out GRPO Signal: Adaptive Trace Prefix Control for Hard Reasoning Problems
- The Unlearnability Phenomenon in RLVR for Language Models
- Training Verifiers to Solve Math Word Problems
- The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models
- Harder Is Better: Boosting Mathematical Reasoning via Difficulty-Aware GRPO and Multi-Aspect Question Reformulation
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Measuring Mathematical Problem Solving With the MATH Dataset
- Tulu 3: Pushing Frontiers in Open Language Model Post-Training
- Enhancing Multi-Modal LLMs Reasoning via Difficulty-Aware Group Normalization
- Understanding R1-Zero-Like Training: A Critical Perspective
- DRIFT: Difficulty Routing Self-DIstillation with Rhythm-Gated Exploration and Success BuFfer Training
- Beyond Variance: Prompt-Efficient RLVR via Rare-Event Amplification and Bidirectional Pairing
- Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Models
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Solving math word problems with process- and outcome-based feedback
- Qwen2.5 Technical Report
- Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Training
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems