dataset_path: sjyuxyz/MMLU-Pro-with-subset
test_split: test
fewshot_split: dev
fewshot_config:
  sampler: first_n
output_type: generate_until
doc_to_text: "{% set alphabet = 'ABCDEFGHIJ' %}{{ question.strip() }}{% for index in range(options|length) %}\n{{ alphabet[index] }}. {{ options[index] }}{% endfor %}\nAnswer:"
doc_to_target: "{{['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J'][answer_index]}}"
generation_kwargs:
  until:
    - "</s>"
    - "\n"
metric_list:
  - metric: exact_match
    aggregation: mean
    higher_is_better: true
metadata:
  version: 1.0
dataset_kwargs:
  trust_remote_code: true
