GSM8K

Dataset Card for GSM8K Dataset Summary GSM8K (Grade School Math 8K) is a dataset of 8.5K high quality linguistically diverse grade school math word problems. The dataset was created to support the task of question answering on basic mathematical problems that require multi-step reasoning. These problems take between 2

Papers using GSM8K (53)

Stable Reinforcement Learning for Efficient Reasoning2025 · 25 citesThink or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models2025 · 17 citesReasoning Under 1 Billion: Memory-Augmented Reinforcement Learning for Large Language Models2025 · 7 citesQeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMs2025 · 4 citesOpen-Medical-R1: How to Choose Data for RLVR Training at Medicine Domain2025 · 3 citesMOTIF: Modular Thinking via Reinforcement Fine-tuning in LLMs2025 · 2 citesEnhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization2024 · 2 citesDiscrete Tilt Matching2026Boosting Accuracy and Efficiency of Budget Forcing in LLMs via Reinforcement Learning for Mathematical Reasoning2025 · 1 citesEntropy-Regularized Process Reward Model2024 · 1 citesBeyond KL Divergence: Policy Optimization With Flexible Bregman Divergences For LLM Reasoning2026ESSAM: A Novel Competitive Evolution Strategies Approach to Reinforcement Learning for Memory Efficient LLMs Fine-Tuning2026Think Dense, Not Long: Dynamic Decoupled Conditional Advantage for Efficient Reasoning2026TMS: Trajectory-Mixed Supervision for Reward-Free, On-Policy SFT2026$n$-Musketeers: Reinforcement Learning Shapes Collaboration Among Language Models2026UpSkill: Mutual Information Skill Learning for Structured Response Diversity in LLMs2026Why GRPO Needs Normalization: A Local-Curvature Perspective on Adaptive Gradients2026Group-Aware Reinforcement Learning for Output Diversity in Large Language Models2025 · 1 citesScRPO: From Errors to Insights2025h1: Bootstrapping LLMs to Reason over Longer Horizons via Reinforcement Learning2025GIFT: Group-Relative Implicit Fine-Tuning Integrates GRPO with DPO and UNA2025It's Not You, It's Clipping: A Soft Trust-Region via Probability Smoothing for LLM RL2025wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models2025Efficient Online RFT with Plug-and-Play LLM Judges: Unlocking State-of-the-Art Performance2025RL in Name Only? Analyzing the Structural Assumptions in RL post-training for LLMs2025Maximizing Confidence Alone Improves Reasoning2025Segment Policy Optimization: Effective Segment-Level Credit Assignment in RL for Large Language Models2025Synthetic Data Generation & Multi-Step RL for Reasoning & Tool Use2025Training Large Language Models to Reason via EM Policy Gradient2025Tapered Off-Policy REINFORCE: Stable and efficient reinforcement learning for LLMs2025Coevolving with the Other You: Fine-Tuning LLM with Sequential Cooperative Multi-Agent Reinforcement Learning2024 · 3 citesQwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement2024 · 12 citesRUMAD: Reinforcement-Unifying Multi-Agent Debate2026 · 2 citesVinePPO: Refining Credit Assignment in RL Training of LLMs2024 · 1 citesOffline Reinforcement Learning for LLM Multi-Step Reasoning2024 · 1 citesCPL: Critical Plan Step Learning Boosts LLM Generalization in Reasoning Tasks2024SMART: Self-learning Meta-strategy Agent for Reasoning Tasks2024Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models2024Big-Math: A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models2025Towards Hierarchical Multi-Step Reward Models for Enhanced Reasoning in Large Language Models2025TutorGym: A Testbed for Evaluating AI Agents as Tutors and Students2025Inpainting-Guided Policy Optimization for Diffusion Large Language Models2025Reward Model Routing in Alignment2025The Good, The Bad, and The Hybrid: A Reward Structure Showdown in Reasoning Models Training2025Semantic Soft Bootstrapping: Long Context Reasoning in LLMs without Reinforcement Learning2025Differentiable Evolutionary Reinforcement Learning2025AGGC: Adaptive Group Gradient Clipping for Stabilizing Large Language Model Training2026SD-E$^2$: Semantic Exploration for Reasoning Under Token Budgets2026Beyond KL Divergence: Policy Optimization with Flexible Bregman Divergences for LLM Reasoning2026Don't Overthink It: Inter-Rollout Action Agreement as a Free Adaptive-Compute Signal for LLM Agents2026Synthetic Data RL: Task Definition Is All You Need2025RL for Reasoning by Adaptively Revealing Rationales2025MASPRM: Multi-Agent System Process Reward Model2025