The pre-trained Large Language Models (LLMs) can be adapted for many downstream tasks and tailored to align with human preferences through fine-tuning. Recent studies have discovered that LLMs can achieve desirable performance with only a small amount of high-quality data, suggesting that a large amount of the data in these extensive datasets is redundant or even harmful. Identifying high-quality data from vast datasets to curate small yet effective datasets has emerged as a critical challenge. In this paper, we introduce SHED, an automated dataset refinement framework based on Shapley value for instruction fine-tuning. SHED eliminates the need for human intervention or the use of commercial LLMs. Moreover, the datasets curated through SHED exhibit transferability, indicating they can be reused across different LLMs with consistently high performance. We conduct extensive experiments to evaluate the datasets curated by SHED. The results demonstrate SHED's superiority over state-of-the-art methods across various tasks and LLMs; notably, datasets comprising only 10% of the original data selected by SHED achieve performance comparable to or surpassing that of the full datasets.
Hong Kong Metropolitan University(香港都会大学);Hong Kong Polytechnic University(香港理工大学);Shanghai 100me Internet Technology Co., Ltd.(上海一零零米互联网科技有限公司);Lingnan University(岭南大学)
How Likely Do LLMs with CoT Mimic Human Reasoning?
Zhejiang University(浙江大学);School of Engineering, Westlake University(西湖大学工学院);Institute of Advanced Technology, Westlake Institute for Advanced Study(西湖高等研究院先进技术研究院)
GeoGalactica: A Scientific Large Language Model in Geoscience
Shanghai Jiao Tong University(上海交通大学);University of Michigan(密歇根大学);Institute of Geographical Science and Natural Resources Research, CAS(中国科学院地理科学与资源研究所);The Hong Kong University of Science and Technology(香港科技大学)
EgoPlan-Bench: Benchmarking Multimodal Large Language Models for Human-Level Planning
Tencent AI Lab(腾讯 AI 实验室);The University of Hong Kong(香港大学);ARC Lab, Tencent PCG(腾讯平台与内容事业群 ARC 实验室);University of California, Berkeley(加州大学伯克利分校);Peng Cheng Laboratory(鹏城实验室)
Department of Computer Science and Technology, Tsinghua University(清华大学计算机科学与技术系);Beijing National Research Center for Information Science and Technology(北京信息科学与技术国家研究中心);Institute of Trustworthy Embodied AI, Fudan University(复旦大学可信具身人工智能研究院)
ProMQA: Question Answering Dataset for Multimodal Procedural Activity Understanding
Language Technologies Institute, Carnegie Mellon University(语言技术研究所,卡内基梅隆大学);National Institute of Advanced Industrial Science and Technology (AIST)(国家先进工业科学与技术研究院)
Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement
Massachusetts Institute of Technology(麻省理工学院);Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of Washington(华盛顿大学);University of Southern California(南加州大学)