A judge scoring F1 0.000 at the default threshold scores 0.860 tuned. Same judge, same scores. A teaching library and lab course for LLM and agent evaluation. Pure Python standard library — no ...
Some results have been hidden because they may be inaccessible to you
