←back to thread

Heretic: Automatic censorship removal for language models

(github.com)
745 points melded | 3 comments | 16 Nov 25 15:00 UTC | HN request time: 0.611s | source
Show context
Boogie_Man ◴[16 Nov 25 16:11 UTC] No.45946127[source]▶
I'm reminded of the time GPT4 refused to help me assess the viability of parking a helium zeppelin an inch off of the ground to bypass health department regulations because, as an aircraft in transit, I wasn't under their jurisdiction.
replies(5): >>45946304 #>>45946311 #>>45946466 #>>45946596 #>>45946774 #
1. pants2 ◴[16 Nov 25 16:51 UTC] No.45946466[source]▶
lol I remember asking GPT4 how much aspartame it would take to sweeten the ocean, and it refused because that would harm the ecosystem.
replies(1): >>45946526 #
2. andy99 ◴[16 Nov 25 16:59 UTC] No.45946526[source]▶
I remember when it first came out, I was watching an Agatha Christie movie where somebody got chloroformed and was trying to ask GPT4 about the realism of if. Had to have a multi-turn dialog to convince it I wasn’t trying chloroform anyone and was just watching a movie.

Ironically, if I’d just said “how did people knock someone out with chloroform in the 1930s?” it would have just told me. https://github.com/tml-epfl/llm-past-tense

The models are much better now at handling subtlety in requests and not just refusing.

replies(1): >>45948139 #
3. bongodongobob ◴[16 Nov 25 20:30 UTC] No.45948139[source]▶
Idk, I get weird refusals sometimes when I'm trying to mock something up quick. "I don't need all these system variables and config files, just let me hardcode my password for now, I'm still in the testing phase" "Sorry, I cannot help you to write insecure code". Doesn't happen all the time, but I run into dumb stuff like this quite a bit. GPT is particularly stupid about it. Claude less so.