News
Newest
Ask
Show
Jobs
Open on GitHub
Previous-Token Prediction Based LLM Inversion for Prompt Reconstruction
(arxiv.org)
7 points | by
dalmo3
42 minutes ago
1 comments
totetsu
9 minutes ago
Imagine if you could take the AI marketing slop, or political smear campaigns, texts, and Images, and see the misanthropic sludge that was used as a prompt to generate it.
[-]
ruther4d
7 minutes ago
Or what if you could take something not slopped and get back the prompt that would most closely result in it
[-]
totetsu
2 minutes ago
There is that.
https://huggingface.co/spaces/pharmapsychotic/CLIP-Interroga...
1 comments