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.
Cosmopedia v2, tokenized with GPT-2 (yam)
Cosmopedia v2 (Hugging Face SmolLM corpus, cosmopedia-v2 subset; ODC-By 1.0): synthetic textbooks, stories and lessons generated by Mixtral. This is a sample: whole parquet files picked in seeded random order (list in source_files.json). Please credit the original dataset when using this data.
- Format: uint16 little-endian GPT-2 tokens, no header;
<|endoftext|>precedes each document. - Train: 3,500,000,000 tokens in 35 shards (
train_*.bin); validation: 10,000,000 tokens (val_000000.bin, the first documents of the sample).
Per-shard token counts are in meta.json.
- Downloads last month
- 206