Semanlink - Sanjeev Arora sur Twitter : "A priori, fine-tuning a huge LM on a few datapoints could lead to catastrophic overfitting. So why doesn’t it? Our theory + experiments..."

Impression

Recherche de Mot-clé

Recherche de Doc

Préférences...

Sanjeev Arora sur Twitter : "A priori, fine-tuning a huge LM on a few datapoints could lead to catastrophic overfitting. So why doesn’t it? Our theory + experiments..."

Tags:

Au sujet de ce document

sl:bookmarkOf : https://twitter.com/prfsanjeevarora/status/1580892095757230081?s=20&t=QjjqB8m_xtc1S-4DXLFByg
sl:creationDate : 2022-10-14
sl:creationTime : 2022-10-14T15:06:32Z

Infos sur le fichier

Bookmark of: https://twitter.com/prfsanjeevarora/status/1580892095757230081?s=20&t=QjjqB8m_xtc1S-4DXLFByg

Documents with similar tags (experimental)

Sanjeev Arora sur Twitter : "new `skills' induced by LLM fine-tuning can be localized in tiny fraction of the model."

Tags:

2023-07-07 A propos

Sanjeev Arora sur Twitter : "Fine-tuning language models using just forward pass!...r

Tags:

2023-06-09 A propos