Arena-Hard

The 'Arena-Hard' dataset is a benchmark used to evaluate the performance of reasoning models by assessing their ability to generate outputs that can deceive other LLM judges.

Papers using Arena-Hard (24)

Alignment through Meta-Weighted Online Sampling: Bridging the Gap between Data Generation and Preference Optimization2025Token-weighted Direct Preference Optimization with Attention2026MMoA: An AI-Agent framework with recurrence for Memoried Mixure-of-Agent2026Examining Reasoning LLMs-as-Judges in Non-Verifiable LLM Post-Training2026References Improve LLM Alignment in Non-Verifiable Domains2026The Art of Asking: Multilingual Prompt Optimization for Synthetic Data2025Icon$^{2}$: Aligning Large Language Models Using Self-Synthetic Preference Data via Inherent Regulation2025Not All Preferences are What You Need for Post-Training: Selective Alignment Strategy for Preference Optimization2025P3: Prompts Promote Prompting2025SGPO: Self-Generated Preference Optimization based on Self-Improver2025Robust Preference Optimization via Dynamic Target Margins2025ConfPO: Exploiting Policy Model Confidence for Critical Token Selection in Preference Optimization2025FuseChat-3.0: Preference Optimization Meets Heterogeneous Model Fusion2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models2025Tower+: Bridging Generality and Translation Specialization in Multilingual LLMs2025MaPPO: Maximum a Posteriori Preference Optimization with Prior Knowledge2025Alignment through Meta-Weighted Online Sampling: Bridging the Gap between Data Generation and Preference Optimization2025Exploring the Potential of Offline RL for Reasoning in LLMs: A Preliminary Study2025ComPO: Preference Alignment via Comparison Oracles2025RSPO: Regularized Self-Play Alignment of Large Language Models2025CrowdSelect: Synthetic Instruction Data Selection with Multi-LLM Wisdom2025Capturing Nuanced Preferences: Preference-Aligned Distillation for Small Language Models2025T-REG: Preference Optimization with Token-Level Reward Regularization2024NILE: Internal Consistency Alignment in Large Language Models2024