Dataset Viewer
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Dolma 3 Mix (150B sample), exact-deduplicated (audit corpus)
Audit corpus for "Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior".
Built from the full allenai/dolma3_mix-150B-1025 release by keeping the first occurrence of each document id
(the release repeats high-quality documents under the same id), then tokenizing with the SmolLM3 tokenizer
into DataTrove tokenized-bytes format (128 shuffled shards).
Script: preprocessing/src/tokenize_olmo3_dedup.py in https://github.com/gucci-j/verify-ppt-at-scale
- Downloads last month
- 152