Skip to content

Add NLG metrics to aci_bench#107

Open
jbdel wants to merge 1 commit intoMedARC-AI:mainfrom
jbdel:aci-bench-enhancements
Open

Add NLG metrics to aci_bench#107
jbdel wants to merge 1 commit intoMedARC-AI:mainfrom
jbdel:aci-bench-enhancements

Conversation

@jbdel
Copy link
Contributor

@jbdel jbdel commented Jan 22, 2026

Add automatic NLG metrics (BLEU with smoothing, ROUGE, BERTScore) to the aci_bench environment.

=== Example ===
Judge Scores: accuracy 5/5, completeness 4/5, clarity 5/5

Auto Metrics (NLG):
  bleu: 0.1296
  rouge1: 0.4720
  rougeL: 0.2315
  bertscore_f1: 0.8367

- Add automatic metrics computation with smoothed BLEU
- Fix extra_body parameter issue for OpenAI judge calls
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant