matt-seb-ho/WikiWhy

WikiWhy is a new benchmark for evaluating LLMs' ability to explain between cause-effect relationships. It is a QA dataset containing 9000+ "why" question-answer-rationale triplets.

语言：Python · GitHub 星标：49 · Fork：1

话题：artificial-intelligence · dataset · explainable-ai · iclr2023 · machine-learning · nlp · nlp-datasets · open-domain-qa

在 GitHub 上查看原仓库