README.md

February 19, 2024 ยท View on GitHub

AlpacaEval with Fine-grained pairwise evaluation

Template for GPT-as-Judge

See eval_template_pairwise.md.

Run evaluation in parallel

bash evaluate/eval.sh urial-llama-70b 
bash evaluate/eval.sh urial-llama-7b
bash evaluate/eval.sh mistral-urial

Contents

  1. 1AlpacaEval with Fine-grained pairwise evaluation
  2. 2Template for GPT-as-Judge
  3. 3Run evaluation in parallel