hkust-nlp/PreSelect-100B
๐ Paper | ๐จ fastText Classifier | ๐ค Released Dataset | ๐ฆ Repo PreSelect-100B is a curated ~100B token pretraining dataset that achieves great performance on various benchmarks. It is filtered by PreSelect-Classifier at 10% threshold, where the pool is a randomly sampled subset of DCLM-refinedweb, which is a cleaned version of Common Crawl raw data but without any model-based filtering.โฆ See the full description on the dataset page: https://huggingface.co/datasets/hkust-nlp/PreSelect-100B.
| Author / organization | hkust-nlp |
|---|---|
| License (as declared) | mit |
| Downloads | 408 (as reported by Hugging Face) |
| Likes | 11 |
| Created | 18 February 2025 |
| Last updated at source | 4 March 2025 |
Popularity, last 3 days
408 now (+23 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Recent changes
- 16 Oct 2026 Description: ๐ Paper | ๐จ fastText Classifier | ๐ค Released Dataset | ๐ฆ Repo PreSelect-100B is a curated ~100B token pretraining dataset that achieves great performance on various benchmarks. It is filtered by PreSelect-Classifier at 10% threshold, wh โ ๐ Paper | ๐จ fastText Classifier | ๐ค Released Dataset | ๐ฆ Repo PreSelect-100B is a curated ~100B token pretraining dataset that achieves great performance on v
Changes detected by Silytics between refreshes of the source.
More from hkust-nlp
- hkust-nlp/SimpleRL-Zoo-DataDataset
- hkust-nlp/dart-math-hardDataset
- hkust-nlp/CodeIO-PyEdu-Reasoning-RawDataset
- hkust-nlp/dart-math-uniformDataset
- hkust-nlp/agentboardDataset
- hkust-nlp/CodeIO-PyEdu-ReasoningDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.