{
  "dataset_id": "tau/commonsense_qa",
  "train_split": "train",
  "eval_split": "validation",
  "train_examples": 2000,
  "eval_examples": 1221,
  "skipped_without_answer": 0,
  "dropped_overlength_train": 0,
  "dropped_overlength_eval": 0,
  "max_length": 512,
  "completion_only_loss": true,
  "prompt_template": "commonsenseqa_letter"
}