tag: - math_word_problems task: gsm8k_cot_zeroshot dataset_path: gsm8k dataset_name: main output_type: generate_until training_split: train fewshot_split: train test_split: test process_results: !function utils.process_results doc_to_text: "Q: {{question}}\nA: Let's think step by step." doc_to_target: "{{answer}}" #" {{answer.split('### ')[-1].rstrip()}}" metric_list: - metric: math_verify aggregation: mean higher_is_better: true generation_kwargs: until: - "Q:" - "" - "<|im_end|>" do_sample: false repeats: 1 num_fewshot: 0 metadata: version: 3.0