HuggingFaceFW/fineweb
FineWeb is a dataset of over 15 trillion tokens of cleaned and deduplicated English web data from CommonCrawl. It is optimized for LLM performance and processed using the datatrove library. The dataset aims to provide high-quality data for training large language models and outperforms other commonly used web datasets.We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Detaljeret Introduktion
FineWeb is a large-scale dataset designed to provide high-quality web data for training large language models. It includes over 15 trillion tokens of cleaned and deduplicated English web data from CommonCrawl. The dataset is processed using the datatrove library and is optimized for LLM performance. It outperforms other commonly used web datasets in benchmark tasks.
MereDatasæt
Ithaka 2006 Survey of US Higher Education Faculty Attitudes and Behaviors
This study surveys the attitudes and behaviors of US higher education faculty members regarding online resources, the library, and related topics. It covers a wide range of issues, including faculty dependence on electronic scholarly resources, the transition from print to electronic journals, publishing preferences, e-books, and the preservation of scholarly journals.
ToM QA Dataset: Evaluating Theory of Mind in Question Answering
The ToM QA Dataset is designed to evaluate question-answering models' ability to reason about beliefs. It includes 3 task types and 4 question types, creating 12 total scenarios. The dataset is inspired by theory-of-mind experiments in developmental psychology and is used to test models' understanding of beliefs and inconsistent states of the world.
Psy-Insight: Mental Health Counseling Dataset
Psy-Insight is a bilingual, interpretable multi-turn dataset for mental health counseling dialogues. It includes 6,208 rounds of multi-turn counseling dialogues in English and 5,776 rounds in Chinese, annotated with step-by-step reasoning labels and multi-task labels. This dataset is designed to support the application of large language models in mental health and is suitable for tasks such as emotion classification and psychological treatment interpretation.