task: piqa_generation dataset_path: piqa dataset_name: null output_type: generate_until training_split: train validation_split: validation test_split: null doc_to_text: |- Given the following goal and two candidate answers (A and B), choose the best answer Goal: {{goal}} A. {{sol1}} B. {{sol2}} Your response should end with "The best answer is [the_answer_letter]" where the [the_answer_letter] is one of A or B. doc_to_target: "['A', 'B'][label]" doc_to_decontamination_query: goal metric_list: - metric: exact_match aggregation: mean higher_is_better: true metadata: version: 1.0 dataset_kwargs: trust_remote_code: true