A High-Quality Benchmark for Portuguese Academic Reasoning in Large Language Models
ENEM 2025 Dataset is a curated collection of question-answer pairs derived from the 2025 edition of the Brazilian National High School Exam (ENEM), designed to evaluate and improve the reasoning, reading comprehension, and multiple-choice answering capabilities of large language models in Brazilian Portuguese; as one of the largest… See the full description on the dataset page:
https://huggingface.co/datasets/sapiens-technology/enem_2025.
