BSC-LT/IFEval_es

IFEval_es（西班牙语指令跟随评估基准）是一个西班牙语的提示数据集，专业翻译自英文版本的IFEval数据集。该数据集旨在评估聊天或指令微调语言模型的能力，包含541个可验证的指令，如“写超过400字”和“至少提及AI关键词3次”，这些指令可以通过启发式方法进行验证。每个实例仅包含一个输入提示。数据集的结构为JSONL格式，每个行包含实例标识符和相应的输入提示。数据集的创建动机是为了解决大语言模型指令跟随能力评估的标准化问题，并通过翻译来提升西班牙语在NLP领域的支持。数据集的翻译过程遵循了特定的指南，以确保翻译的准确性和一致性。

IFEval_es (Instruction-Following Eval benchmark - Spanish) is a prompt dataset in Spanish, professionally translated from the main version of the IFEval dataset in English. The dataset is designed to evaluate chat or instruction fine-tuned language models, comprising 541 verifiable instructions such as write in more than 400 words and mention the keyword of AI at least 3 times which can be verified by heuristics. Each instance contains just one input prompt. The dataset is provided in JSONL format, where each row corresponds to a prompt and contains an instance identifier and the corresponding input prompt. The creation rationale for the dataset is to address the standardization of instruction-following ability evaluation in large language models and to improve Spanish support in the NLP field through translation. The translation process followed specific guidelines to ensure accuracy and consistency.