MATH-500

Dataset Card for MATH-500 This dataset contains a subset of 500 problems from the MATH benchmark that OpenAI created in their Let's Verify Step by Step paper. See their GitHub repo for the source file: https://github.com/openai/prm800k/tree/main?tab=readme-ov-file#math-splits

Papers using MATH-500 (15)

Geometric Signatures of Reasoning: A Spectral Perspective on Task Hardness2026SPARK: Susceptibility-Guided Profiling and Steering of Latent Reasoning States in Large Language Models2026TreeThink: A Modular Tree Search Library for Mathematical Reasoning with LLMs2026SLIM-RL: Risk-Budgeted Random-Masking RL for Diffusion LLMs Without Trajectory Slicing2026Maximizing Rollout Informativeness under a Fixed Budget: A Submodular View of Tree Search for Tool-Use Agentic Reinforcement Learning2026SAGE-32B: Agentic Reasoning Via Iterative Distillation2026Counterfactual Credit Policy Optimization for Multi-Agent Collaboration2026What If We Allocate Test-Time Compute Adaptively?2026PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models2026ScRPO: From Errors to Insights2025SIGMA: Search-Augmented On-Demand Knowledge Integration for Agentic Mathematical Reasoning2025From Implicit Exploration to Structured Reasoning: Leveraging Guideline and Refinement for LLMs2025Reinforce LLM Reasoning through Multi-Agent Reflection2025A*-Decoding: Token-Efficient Inference Scaling2025Walk Before You Run! Concise LLM Reasoning via Reinforcement Learning2025