Support mtbench (#5025)

Co-authored-by: Xu Yuanchen <yuanchen.xu00@gmail.com>
2025-09-06 19:40:28 +00:00 · 2023-11-09 13:41:50 +08:00
parent f71e63b0f3
commit 239cd92eff
9 changed files with 312 additions and 13 deletions
--- a/applications/ColossalEval/colossal_eval/evaluate/dataset_evaluator/metrics.py
+++ b/applications/ColossalEval/colossal_eval/evaluate/dataset_evaluator/metrics.py
@@ -185,6 +185,7 @@ metrics4subcategory = {
        "ppl_score_over_choices": ["ALL"],
        "ppl_score": ["ALL"],
    },
+    "mtbench": {"mtbench_single_judge": ["ALL"]},
 }