←back to thread

The Beginner's Guide to Visual Prompt Injections (2023)

(www.lakera.ai)
183 points k5hp | 2 comments | 13 Nov 24 18:07 UTC | HN request time: 0s | source
Show context
driverdan ◴[14 Nov 24 03:39 UTC] No.42132939[source]▶
It looks like this is an old article they changed the date on to get traffic to their site. Image processing was added over a year ago and as someone else mentioned gpt4o responds differently.

It's also strange that they keep referring to "GPT-V4" and in some cases "GPT-4V". OpenAI has never called it V4 (or 4V).

replies(1): >>42133160 #
simonw ◴[14 Nov 24 04:25 UTC] No.42133160[source]▶
OpenAI did call it GPT-4 Vision / GPT-4V when it first came out as a preview last year - I wrote about it at the time: https://simonwillison.net/2023/Oct/14/multi-modal-prompt-inj...

They use that name on the system card here: https://openai.com/index/gpt-4v-system-card/

Looks like you were right about the article being a year old - here is the same article on the Internet archive dated 1st November 2023 https://web.archive.org/web/20231101161710/https://www.laker...

replies(1): >>42136204 #
1. driverdan ◴[14 Nov 24 14:02 UTC] No.42136204[source]▶
I stand corrected. It seems they used it in limited context.
replies(1): >>42146112 #
2. simonw ◴[15 Nov 24 11:56 UTC] No.42146112[source]▶
I don't think they ever made that model available outside of a beta under that name - the first openly available API release of image support was branded GPT-4 Turbo: https://help.openai.com/en/articles/8555496-gpt-4-vision-api