News for hackers. Stuff that matters.

A comment on Bing: “I will not harm you unless you harm me first”. Parent comment

Re:Bing: “I will not harm you unless you harm me first” by simonw on Wednesday February 15, 2023 @11:16AM (#34805886)
That's the big question I have: ChatGPT is way less likely to go into weird threat mode. Did Bing get completely different RLHF, or did they skip that step entirely?
[ Parent | on HN ]
Re:Bing: “I will not harm you unless you harm me first” by jerjerjer on Wednesday February 15, 2023 @11:28AM (#34806091)
Not the first time MS releases an AI into the wild with little guardrails. This is not even the most questionable AI from them.
[ Parent | on HN ]
Re:Bing: “I will not harm you unless you harm me first” by whimsicalism on Wednesday February 15, 2023 @05:51PM (#34812046)
Could just be differences in the prompt.

My guess is that ChatGPT has considerably more RL-HF data points at this point too.

[ Parent | on HN ]
Re:Bing: “I will not harm you unless you harm me first” by moffkalast on Wednesday February 15, 2023 @07:55PM (#34813408)
Speaking of, I wonder if one could get the two to talk to each other by pasting prompts and answers.

Edit: Someone did, they just decided to compare notes lol: https://www.reddit.com/r/bing/comments/112zx36/a_conversatio...

[ Parent | on HN ]