DataRankOpen dataset benchmarks

Dataset Leaderboard

Community-scored datasets for ML pretraining & evaluation.

#DatasetOverallQualityCoverage
Showing 1–5 of 5 datasetsUpdated 2h ago

OpenWebText2

Deduplicated web corpus · 17M documents

Rank #1 overall
Metadata
Creator
EleutherAI
License
CC-BY 4.0
Category
NLP
Size (rows)
17.2M rows
Date range
2023-01-01 → 2024-12-31
Metric Breakdown

Scores blend human review panels and automated quality probes (0–100).

Provenance
Rank History — trailing 12 months
JanAprJulOctJan

Benchmark Comparison

Current dataset vs. its two closest peers across five metrics.

DatasetOverallQualityCoverageFreshnessDiversity

Best score per metric is highlighted. Tap “View Leaderboard” for the full ranking.

Filters applied