MMLU-Pro

MMLU-Pro Dataset MMLU-Pro dataset is a more robust and challenging massive multi-task understanding dataset tailored to more rigorously benchmark large language models' capabilities. This dataset contains 12K complex questions across various disciplines. |Github | 🏆Leaderboard | 📖Paper | 🚀 What's New [2026.03.11] Ad

Papers using MMLU-Pro (17)

Nemotron-CrossThink: Scaling Self-Learning beyond Math Reasoning2025 · 24 citesOpen-Medical-R1: How to Choose Data for RLVR Training at Medicine Domain2025 · 3 citesGroup-Aware Reinforcement Learning for Output Diversity in Large Language Models2025 · 1 citesWarm Up Before You Train: Unlocking General Reasoning in Resource-Constrained Settings2025 · 1 citesVSPO: Vector-Steered Policy Optimization for Behavioral Control2026Apriel-1.5-OpenReasoner: RL Post-Training for General-Purpose and Efficient Reasoning2026SUPERNOVA: Eliciting General Reasoning in LLMs with Reinforcement Learning on Natural Instructions2026Reinforcement Inference: Leveraging Uncertainty for Self-Correcting Language Model Reasoning2026FlipLLM: Efficient Bit-Flip Attacks on Multimodal LLMs using Reinforcement Learning2025TraPO: A Semi-Supervised Reinforcement Learning Framework for Boosting LLM Reasoning2025Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning2025Learning Reasoning Rewards from Expert Demonstrations with Inverse Reinforcement Learning2025Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation2025UR$^2$: Unify RAG and Reasoning through Reinforcement Learning2025Reinforcement Fine-Tuning Naturally Mitigates Forgetting in Continual Post-Training2025Step-wise Policy for Rare-tool Knowledge (SPaRK): Offline RL that Drives Diverse Tool Use in LLMs2025Reinforcing General Reasoning without Verifiers2025