LLM Evaluation Metrics

  • BLEU,
  • ROUGE,
  • perplexity which quantify the similarity between generated text and reference outputs.

LLM