AIME-24

AIME-24 is a benchmark used to evaluate the performance of KV-cache quantization methods in reasoning tasks involving large language models.

Papers using AIME-24 (33)

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning2025 · 1 citesTransformation-Augmented GRPO for Enhancing Exploration in Reasoning of Large Language Models2026LycheeDecode: Accelerating Long-Context LLM Inference via Hybrid-Head Sparse Decoding2026p1: Better Prompt Optimization with Fewer Prompts2026Introspective Diffusion Language Models2026From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning2026Smaller Models are Natural Explorers for Policy-Level Diversity in GRPO2026Off-the-Shelf LLMs as Process Scorers: Training-Free Alternative to PRMs for Mathematical Reasoning2026KVarN: Variance-Normalized KV-Cache Quantization Mitigates Error Accumulation in Reasoning Tasks2026Learn Hard Problems During RL with Reference Guided Fine-tuning2026Benchmarking EngGPT2-16B-A3B against Comparable Italian and International Open-source LLMs2026Test-time Recursive Thinking: Self-Improvement without External Feedback2026Long Chain-of-Thought Compression via Fine-Grained Group Policy Optimization2026MoL-RL: Distilling Multi-Step Environmental Feedback into LLMs for Feedback-Independent Reasoning2025Process Reward Models That Think2025SEED-GRPO: Semantic Entropy Enhanced GRPO for Uncertainty-Aware Policy Optimization2025Skywork Open Reasoner 1 Technical Report2025Inference-Time Hyper-Scaling with KV Cache Compression2025ProtoReasoning: Prototypes as the Foundation for Generalizable Reasoning in LLMs2025Beyond Pass@1: Self-Play with Variational Problem Synthesis Sustains RLVR2025DCPO: Dynamic Clipping Policy Optimization2025SimpleTIR: End-to-End Reinforcement Learning for Multi-Turn Tool-Integrated Reasoning2025Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling2025Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond2025Reinforcement Learning for Reasoning in Small LLMs: What Works and What Doesn't2025First Finish Search: Efficient Test-Time Scaling in Large Language Models2025Which Data Attributes Stimulate Math and Code Reasoning? An Investigation via Influence Functions2025PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model Reasoning2025ScaleDiff: Scaling Difficult Problems for Advanced Mathematical Reasoning2025Skill-Targeted Adaptive Training2025Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning2025Scaling Reasoning without Attention2025SRPO: A Cross-Domain Implementation of Large-Scale Reinforcement Learning on LLM2025