{ "name": "ENEM", "release_date": "2024-11-01", "subsets": { "main": { "language": [ "pt" ], "modalities": [ "multi_image_start", "text" ], "task_type": "multiple_choice_vqa", "prompt_template": "Eu vou te passar uma questão de múltipla escolha. Selecione a alternativa correta (\"A.\", \"B.\", \"C.\", \"D.\" ou \"E.\").\n{% for _ in range(n_figures) %}{% endfor %}{{ question | replace('[[placeholder]]', '') }}\n{% for k, v in options.items() %}{{ k }}. {{ v }}\n{% endfor %}Resposta:", "mapping_from_source": { "media": { "from": "figures", "type": "list", "min_items": 0, "max_items": 6 }, "id": { "from": "id" }, "question": { "from": "question" }, "answer": { "from": "label", "optional": true }, "options": { "from": "alternatives", "optional": true, "note": "list source values are normalized to {A,B,...} dict" }, "extra": { "n_figures": { "from": "n_figures" }, "exam": { "from": "exam" } }, "source": { "format": "json", "url": { "2022": "https://huggingface.co/datasets/maritaca-ai/enem", "2023": "https://huggingface.co/datasets/maritaca-ai/enem", "2024": "https://huggingface.co/datasets/maritaca-ai/enem" } } }, "prompt_template_source": { "origin": "official", "reference": "https://github.com/piresramon/gpt-4-enem (piresramon/gpt-4-enem description.json + enem_multimodal.py:format_example) — Portuguese ENEM prompt as published by the dataset authors", "notes": "Tier 1: ENEM official Portuguese prompt as published by piresramon/gpt-4-enem (the dataset's official evaluation harness)." } } } }