diff --git a/evals/registry/evals/hanabi_clues.yaml b/evals/registry/evals/hanabi_clues.yaml index 977a02ccc9..4218b8103b 100644 --- a/evals/registry/evals/hanabi_clues.yaml +++ b/evals/registry/evals/hanabi_clues.yaml @@ -1,5 +1,5 @@ hanabi_clues: - description: Test the model's ability to list all valid color and numerical clues for a Hanabi hand + description: Test the model's ability to list all valid color and numerical clues for a hand in the game Hanabi, while (in some variants) not listing any invalid clues id: hanabi_clues-list_all_valid_color_and_number.full-v0 metrics: - accuracy