HuggingFaceFW/fineweb-2
FineWeb-2 is a dataset of over 15 trillion tokens of cleaned and deduplicated English web data from CommonCrawl. This is the second iteration of the popular 🍷 FineWeb dataset, bringing high quality pretraining data to over 1000 🗣️ languages.The 🥂 FineWeb2 dataset is fully reproducible, available under the permissive ODC-By 1.0 license and extensively validated through hundreds of ablation experiments.In particular, on the set of 9 diverse languages we used to guide our processing decisions, 🥂 FineWeb2 outperforms other popular pretraining datasets covering multiple languages (such as CC-100, mC4, CulturaX or HPLT, while being substantially larger) and, in some cases, even performs better than some datasets specifically curated for a single one of these languages, in our diverse set of carefully selected evaluation tasks: FineTasks.
Detailed Introduction
FineWeb-2 is a large-scale dataset designed to provide high-quality web data for training large language models. This is the second iteration of the popular 🍷 FineWeb dataset, bringing high quality pretraining data to over 1000 🗣️ languages.The 🥂 FineWeb2 dataset is fully reproducible, available under the permissive ODC-By 1.0 license and extensively validated through hundreds of ablation experiments.In particular, on the set of 9 diverse languages we used to guide our processing decisions, 🥂 FineWeb2 outperforms other popular pretraining datasets covering multiple languages (such as CC-100, mC4, CulturaX or HPLT, while being substantially larger) and, in some cases, even performs better than some datasets specifically curated for a single one of these languages, in our diverse set of carefully selected evaluation tasks: FineTasks.
MoreDataset
t4d: Conversion Algorithm from ToMi to T4D Dataset
This project implements the conversion algorithm from the ToMi dataset to the T4D (Thinking is for Doing) dataset, as introduced in the paper https://arxiv.org/abs/2310.03051. It filters examples with Theory of Mind (ToM) questions and adapts the algorithm to account for second-order false beliefs.
Adolescent Mental Health: WHO Report
The WHO report on adolescent mental health describes actions undertaken by international development organizations to address adolescents’ mental health needs at the country level. It highlights the inadequacy of current efforts and the need for more coordinated and comprehensive interventions.
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.
