AlpacaEval 2.0

'AlpacaEval 2.0' is a dataset/benchmark used to evaluate the alignment of Large Language Models (LLMs) with human preferences through preference optimization methods.

Papers using AlpacaEval 2.0 (19)

Label-Free Reinforcement Learning via Cross-Model Entropy2026TACOS: Open Tagging and Comparative Scoring for Instruction Fine-Tuning Data Selection2025 · 1 citesTransitivity Meets Cyclicity: Explicit Preference Decomposition for Dynamic Large Language Model Alignment2026MMoA: An AI-Agent framework with recurrence for Memoried Mixure-of-Agent2026S-SPPO: Semantic-Calibrated Self-Play Preference Optimization2026C2: Scalable Rubric-Augmented Reward Modeling from Binary Preferences2026Aligning Large Language Models via Fully Self-Synthetic Data2025Icon$^{2}$: Aligning Large Language Models Using Self-Synthetic Preference Data via Inherent Regulation2025SGPO: Self-Generated Preference Optimization based on Self-Improver2025Unlocking Recursive Thinking of LLMs: Alignment via Refinement2025Pre-DPO: Improving Data Utilization in Direct Preference Optimization Using a Guiding Reference Model2025Temporal Self-Rewarding Language Models: Decoupling Chosen-Rejected via Past-Future2025MaPPO: Maximum a Posteriori Preference Optimization with Prior Knowledge2025This Is Your Doge, If It Please You: Exploring Deception and Robustness in Mixture of LLMs2025Rethinking Mixture-of-Agents: Is Mixing Different Large Language Models Beneficial?2025Beyond Sample-Level Feedback: Using Reference-Level Feedback to Guide Data Synthesis2025FocalPO: Enhancing Preference Optimizing by Focusing on Correct Preference Rankings2025AIPO: Improving Training Objective for Iterative Preference Optimization2024Self-Boosting Large Language Models with Synthetic Preference Data2024