A packed pretraining corpus with 100,000,000,000 training input positions and 10,000,000 held-out validation positions, prepared for the recurrent Purrence Hyperbullet language model. It uses a frozen 2,048-ID tokenizer (2,047 ordinary IDs plus structural boundary ID 2,047). Upload status: consider this repository complete only when UPLOADCOMPLETE.json is present. That receipt identifies the immutable data revision whose remote file sizes and checksums were verified. This is a custom memory-mapped corpus, not a Parquet dataset for datasets.loaddataset. Download the core training files with the Hugging Face Hub client: For reproducibility, pass the revision from UPLOADCOMPLETE.json after…
Independent publisher
Jon Su
Jonnester
Models in Library0
Datasets in Library1
Models on Hugging Face84
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Text generation.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.