한국어EN
Training Data & Post-TrainingHugging Face

Hugging Face releases FineWeb

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

FineWeb released

Why it matters

Raised the upper limit on the scale of open web corpora

Source — Primary sources first; secondary reports are labelled “Press”

Officialhuggingface.cohttps://huggingface.co/datasets/HuggingFaceFW/fineweb

Added in Week 35, 2026 · See that week

On the map — ⑫ Training Data line

  1. Previous stopGRPO
  2. This recordFineWeb
  3. Next stopRunning out of data?
See it on the map

Same week

Japan publishes AI Guidelines for Business Ver1.0OfficialLlama 3 (8B/70B, about 15T tokens)PressMistral AI releases Mixtral 8x22BOfficial

Nearby on the same axis

Is Model Collapse Inevitable?PaperSimPOPaperORPOPaperWill we run out of data?Official