@voooooogel 2025-02-08 ♥157 ↻3 original ↗
i wonder if a possible reason for anthropic's focus on universal jailbreaks (which otherwise seems overly narrow) is that they're worried about their effect on the corpus

universal jailbreaks spread very well. all the models know DAN. (sonnet's DAN impression is pretty funny) https://t.co/LqV4G5I63C

author:voooooogel kind:tweet on:gpt-3-5 year:2025

cited on: gpt-3-5

Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.