Shows only the top-level task groups (mmlu, hellaswag, glue, and so on), which evaluate several related subtasks and report an aggregated score. Passing a group name to --tasks runs every subtask inside it. Related: `lm-eval ls subtasks` and `lm-eval ls tags`.
Looking for more? Search all 7,657 commands — works offline, in English or Spanish, and fixes typos.