3 of 4 in OpenAI × Books1/Books2
2018 / 2020
Public
Brown et al. · 2020-05
Language Models are Few-Shot Learners
“Two internet-based books corpora (Books1 and Books2) appear in the GPT-3 training mix as ordinary data sources.”
Books is a genre. LibGen is a library with a legal history. The paper needed the first word. The disk had the second.
The tell · If a training-set name gets quieter between the folder and the paper, ask what the folder was called.
How this became public: Described in the authors’ MDL privilege opinions.
Keyboard: J / K or ← / → to move. S for sources. Esc leaves the pair.
Pipe: Court. Described in the authors’ MDL privilege opinions.
Public claim
Brown et al., GPT-3 paper authors
2020-05 · Language Models are Few-Shot Learners
Article
Private artifact
OpenAI internal naming, later described in court, research ops
circa 2018-2020 · Internal dataset names
Docket
Default is stay. Originals open in an overlay when we can embed them. A new tab only happens if you ask.