danbraunai/pile-uncopyrighted-tok-shuffled

Pile Uncopyrighted (Tokenized + Shuffled) Globally shuffled version of the Pile (uncopyrighted subset), tokenized into fixed-length sequences of 513 token IDs using EleutherAI/gpt-neox-20b. Each row contains a single input_ids column (513 int32 values). Documents are concatenated with EOS tokens between them, then reshaped into fixed-length sequences (following the TransformerLens approach). Sequences are then globally shuffled (seed=42) so that consecutive rows are not from the… See the full description on the dataset page: https://huggingface.co/datasets/danbraunai/pile-uncopyrighted-tok-shu

<iframe src="https://silytics.si/embed/item/f2c17944-348b-4b2c-b34d-c421182b8d94" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

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 Information may be incomplete or out of date.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.