model_setting: openai_batched_multichoice_v2
max_context_length: 128
max_tokens_instruction: 2056
max_tokens_response: 2056
top_p: 0.9
num_iterations: 1
temperature: 0.7
max_documents_per_subject: 200
max_contexts_per_subject: 2500
icl_examples: 3
icl_dataset: lukaemon/mmlu
icl_split: validation
icl_use_options: True