evalstate/tiny-memorization-synthetic download history

evalstate/tiny-memorization-synthetic is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 37 times (9 in the last 7 days), and 237 times in total. It ranks #335,272 among datasets by monthly downloads.

Tiny memorization synthetic datasets Uniformly-random token datasets replicating the methodology of "How much do language models memorize?" (arXiv:2505.24832). Each variant /<tag>/dataset_tokens.npy is an (N, 65) int64 array: a BOS token (id 2048) followed by 64 uniform iid tokens in

Open evalstate/tiny-memorization-synthetic on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.