aisingapore/Instruction-Following-IFEval

SEA-IFEval是一个用于评估聊天或指令调整的大型语言模型（LLM）能力的基准数据集。它基于IFEval，并由母语人士手动翻译成印度尼西亚语、爪哇语、巽他语、泰语、他加禄语和越南语。该数据集设计用于评估模型遵守提示中提供的约束的能力，例如以特定的词/短语开头或以特定数量的部分回答。

SEA-IFEval is a benchmark dataset for evaluating the capabilities of chat or instruction-tuned large language models (LLMs). It is based on IFEval and has been manually translated by native speakers into Indonesian, Javanese, Sundanese, Thai, Tagalog, and Vietnamese. The dataset is designed to evaluate a models ability to adhere to constraints provided in the prompt, such as beginning a response with a specific word/phrase or answering with a certain number of sections.