Semanlink - Anthropic sur Twitter : "We examine which safety techniques for LMs are more robust to human-written, adversarial inputs ..."

Impression

Recherche de Mot-clé

Recherche de Doc

Préférences...

Anthropic sur Twitter : "We examine which safety techniques for LMs are more robust to human-written, adversarial inputs ..."

Tags:

Au sujet de ce document

sl:bookmarkOf : https://twitter.com/AnthropicAI/status/1562828011505717248?s=20&t=V4E-aE79-FecWG_AKx3KKg
sl:creationDate : 2022-08-25
sl:creationTime : 2022-08-25T18:31:06Z

Infos sur le fichier

Bookmark of: https://twitter.com/AnthropicAI/status/1562828011505717248?s=20&t=V4E-aE79-FecWG_AKx3KKg