INNER CODE UNIT · Python

default

Future-House/BixBench · generate_zeroshot_evals.py:74

        default=-1,
        help="Number of examples to evaluate. Default is -1 for all examples",
    )
    parser.add_argument(
        "--output-dir",
        default="results",
        help="Directory to save results. Default is 'results'",
    )
    parser.add_argument(
        "--output-file", type=str, default=None, help="Output file name (optional)"
    )
    parser.add_argument(
        "--dataset-split",
        default="train",
        help="Dataset split to evaluate. Default is 'test'",
    )
    return parser.parse_args()

View source record →