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.
Math4AI Labs
Small data files for the hands-on labs of Mathematics for AI and Machine Learning (math4ai.org). Each chapter's companion notebook (chapterN-*/code/proj.ipynb) downloads the files it needs from here, pinned to a commit, and checks their SHA-256.
Every file is made by a script in the book's repository (chapterN-*/code/data/make_*.py), which names its sources.
chapter3/
Token embeddings (float16) of the same 2,200 English words, as .npz with arrays words, token_ids, emb (2200 × 768), gender (+1 male, −1 female, 0 none), plural (1, 0, −1 n/a) and occupation. Made by chapter3-subspaces-orthogonality/code/data/make_gpt2_words.py.
| File | Source | License |
|---|---|---|
gpt2_words.npz |
openai-community/gpt2 | MIT |
distilgpt2_words.npz |
distilbert/distilgpt2 | Apache-2.0 |
gpt2austen_words.npz |
GPT-2 fine-tuned by the script for one epoch on three Jane Austen novels (public domain, Project Gutenberg via NLTK) | MIT |
Used by Lab 7 (concept erasure by projection) and Lab 8 (principal angles between subspaces).
- Downloads last month
- 35