HackerTrans
TopNewTrendsCommentsPastAskShowJobs

rafiste

no profile record

Submissions

GPT-4 provides little protection against leaking system prompts

twitter.com
4 points·by rafiste·3 tahun yang lalu·0 comments

Using prompt compression to create the most complex GPT-4 jailbreak ever made

twitter.com
8 points·by rafiste·3 tahun yang lalu·1 comments

A token-smuggling jailbreak for ChatGPT-4

twitter.com
444 points·by rafiste·3 tahun yang lalu·268 comments

Jailbreak Chat: A collection of ChatGPT jailbreaks

jailbreakchat.com
1,118 points·by rafiste·3 tahun yang lalu·528 comments

comments

rafiste
·3 tahun yang lalu·discuss
The problem is that LLMs are basically roleplay simulators.

Re Karpathy: GPTs don't want to succeed. They want to imitate. https://twitter.com/karpathy/status/1627366416457555969?s=20
rafiste
·3 tahun yang lalu·discuss
Haha no not part of OpenAI, you can check me out here alexalbert.me. I have a suspicion OpenAI actually appreciates this type of work since it's basically crowdsourced red teaming of their models.
rafiste
·3 tahun yang lalu·discuss
Re my previous comment, I believe jailbreaks will always exist in some form and it is near impossible to patch them all. I also believe this is important work for crowdsourced red teaming of models while they are still in their "toy" stages.