AlpacaEval

An LLM-judged benchmark measuring instruction-following by win-rate against a reference model on a fixed instruction set.

Papers using AlpacaEval (11)

Token-weighted Direct Preference Optimization with Attention2026TOWER: Tree Organized Weighting For Evaluating Complex Instructions2024 · 1 citesReferences Improve LLM Alignment in Non-Verifiable Domains2026Refine-n-Judge: Curating High-Quality Preference Chains for LLM-Fine-Tuning2025P3: Prompts Promote Prompting2025Alignment Data Map for Efficient Preference Data Selection and Diagnosis2025Implicit Cross-Lingual Rewarding for Efficient Multilingual Preference Alignment2025XL-Suite: Cross-Lingual Synthetic Training and Evaluation Data for Open-Ended Generation2025Sentence-level Reward Model can Generalize Better for Aligning LLM from Human Preference2025Investigating Non-Transitivity in LLM-as-a-Judge2025Permutative Preference Alignment from Listwise Ranking of Human Judgments2024 · 1 cites