LLM-BabyBench-Decompose
收藏arXiv2025-09-30 收录
下载链接：
https://huggingface.co/datasets/salem-mbzuai/LLM-BabyBench
下载链接
链接失效反馈官方服务：
资源简介：
该数据集旨在评估大型语言模型将高级指令分解为连贯子目标序列的能力。它是LLM-BabyBench基准测试套件的一部分，专注于考量大型语言模型在分解任务上的表现。
This dataset is designed to evaluate the ability of large language models (LLMs) to decompose high-level instructions into coherent sequences of sub-goals. As part of the LLM-BabyBench benchmark suite, it focuses on assessing the performance of large language models in task decomposition.
提供机构：
Authors of the paper
