WueNLP/SLAQ

SLAQ（Short-Long Form Alignment for Factual Question Answering）是一个用于评估大型语言模型（LLMs）在事实性问题回答中，短形式问题与长形式问题之间事实一致性及可信度的控制评估框架。该数据集旨在比较当问题以独立（短形式）提出与集成到复杂查询（长形式）中时，LLMs的事实性答案。SLAQ (Short-Long Form Alignment for Factual Question Answering) is a controlled evaluation framework designed to assess the factual consistency and trustworthiness of Large Language Models (LLMs) in factual question answering, comparing their answers when questions are asked in isolation (short-form) versus integrated into complex queries (long-form).