Y Hacker News new | ask | show | jobs

user: cryptohell created: 2023-02-26 karma: 76 submissions: 0 points | 0 comments Can we bootstrap AI Safety despite being unable to even define it? 2 points | 2 comments 0 points | 0 comments Optimality of Frequency Moment Estimation 8 points | 1 comments 0 points | 0 comments LLMs can hide arbitrary undetectable information in their responses 3 points | 3 comments LLMs can hide arbitrary information in their responses, undetectably 1 points | 0 comments Undetectable Watermarks for LLMs 3 points | 0 comments Undetectable Watermarks for LLMs 2 points | 0 comments 0 points | 0 comments Undetectable Watermarks for Language Models 1 points | 1 comments 0 points | 0 comments 0 points | 0 comments Undetectable Watermarks for Language Models 135 points | 69 comments (Provably) Undetectable Watermarks for LLMs 1 points | 0 comments 0 points | 0 comments 0 points | 0 comments