|
- Metas Llama 3. 1 can recall 42 percent of the first Harry . . .
The study authors took 36 books and broke each of them up into overlapping 100-token passages Using the first 50 tokens as a prompt, they calculated the probability that the next 50 tokens will be identical to the original passage
- Meta AI can reproduce half of Harry Potter book verbatim
Meta's artificial intelligence model, Llama 3 1, can recall an impressive 42 percent of the first Harry Potter book, amidst ongoing copyright lawsuits against AI companies
- AI Training and Copyright: Meta’s Llama Model Literally . . .
The study, detailed in a research paper from researchers at Stanford, Cornell, and West Virginia University, found that Meta’s Llama 3 1 70B model could recall a staggering 42% of Harry Potter
- Metas Llama 3. 1 Can Recall 42% of the First Harry Potter Book
Meta's Llama 3 1 can recall 42 percent of the first Harry Potter book New research could have big implications for copyright lawsuits against generative AI Timothy B Lee - Specifically, the paper estimates that Llama 3 1 70B has memorized 42 percent of the first Harry Potter book well enough to reproduce 50-token excerpts at least half the time
- Metas AI memorised books verbatim – that could cost it billions
Such testing revealed that Meta’s Llama 3 1 70B model has memorised most of the first book in J K Rowling’s Harry Potter series, as well as The Great Gatsby and George Orwell’s dystopian
- Who’sHarryPotter? ApproximateUnlearninginLLMs - arXiv. org
We evaluate our technique on the task of unlearning the Harry Potter books from the Llama2-7b model (a generative language model recently open-sourced by Meta) While the model took over 184K GPU-hours to pretrain, we show that in about 1 GPU hour of finetuning, we effectively erase the model’s ability to generate or recall Harry Potter-related
- Who’s Harry Potter? Approximate Unlearning for LLMs
We evaluate our technique on the task of unlearning the Harry Potter books from the Llama2-7b model (a generative language model recently open-sourced by Meta)
|
|
|