HLB
收藏资源简介:
HLB数据集由香港中文大学创建,旨在评估大型语言模型(LLMs)在语言使用中的人类相似性。该数据集包含20个大型语言模型在10个心理语言学实验中的表现,每个实验收集了约50至100条人类参与者的响应和100条LLMs的响应。数据集的创建过程包括使用在线调查平台Qualtrics进行实验设计,并通过自动编码算法提取语言使用模式。HLB数据集主要应用于评估和改进LLMs在自然语言处理中的表现,确保模型能够准确捕捉人类语言的多样性和丰富性。
The HLB dataset was developed by The Chinese University of Hong Kong, with the goal of evaluating the human-likeness of large language models (LLMs) in their language use. This dataset includes the performance data of 20 large language models across 10 psycholinguistic experiments. For each experiment, roughly 50 to 100 responses from human participants and 100 responses from LLMs are collected. The dataset creation process involved designing experiments using the online survey platform Qualtrics and extracting language usage patterns through automatic coding algorithms. The HLB dataset is mainly applied to evaluate and enhance the performance of LLMs in natural language processing, ensuring that the models can accurately capture the diversity and richness of human language.




