lm-evaluation-harness/lm_eval/tasks/bbh/zeroshot/boolean_expressions.yaml

"dataset_name": "boolean_expressions"
"description": "Evaluate the result of a random Boolean expression.\n\n"
"doc_to_text": "Q: {{input}}\nA:"
"include": "_zeroshot_template_yaml"
"task": "bbh_zeroshot_boolean_expressions"

filter_list:
  - name: "strict-match"
    filter:
      - function: "take_first"
  - name: "flexible-extract"
    filter:
      - function: "regex"
        group_select: 0
        regex_pattern: "\\b(True|False)\\b"
      - function: "take_first"