Alpacaeval | search
Click here to view Alpacaeval news from 60+ newspapers.
Bookmark or Share- AlpacaEval: An Automatic Evaluator of Instruction-following ModelsYann Dubois. PhD student in Machine Learning. Follow. Stanford, California; Email; Twitter; LinkedIn; Github; AlpacaEval: An Automatic Evaluator of Instruction ...
- GitHub - tatsu-lab/alpaca_eval: An automatic evaluator for instruction ...AlpacaEval : An Automatic Evaluator for Instruction-following Language Models. AlpacaEval 2.0 with length-controlled win-rates (paper) has a spearman correlation of 0.98 with ChatBot Arena while costing less than $10 of OpenAI credits run and running in less than 3 minutes.
- [2404.04475] Length-Controlled AlpacaEval: A Simple Way to Debias ...As a real case study, we focus on reducing the length bias of AlpacaEval, a fast and affordable benchmark for chat LLMs that uses LLMs to estimate response quality. Despite being highly correlated with human preferences, AlpacaEval is known to favor models that generate longer outputs.
- AlpacaEval Leaderboard - GitHub PagesAlpacaEval is an automatic evaluator for measuring the performance of language models on a task-oriented benchmark. It ranks 50 models by their win rates on length-controlled and community verified datasets, and compares them with GPT-4 as a baseline.
- Releases · tatsu-lab/alpaca_eval - GitHubAn automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast. - tatsu-lab/alpaca_eval
More
