zlacker

>>isp+(OP)
A cautionary tale for why not to put unfiltered ChatGPT output directly to customers.

Nitter mirror: https://nitter.net/ChrisJBakke/status/1736533308849443121

Related - "New kind of resource consumption attack just dropped": https://twitter.com/loganb/status/1736449964006654329 | https://nitter.net/loganb/status/1736449964006654329

>>isp+1
There's no such thing as a filtered LLM output.

How do you plan on avoiding leaks or "side effects" like the tweet here?

If you just look for keywords in the output, I'll ask ChatGPT to encode its answers in base64.

You can literally always bypass any safeguard.

>>iLoveO+13
This is a very good point, and why I would argue that a human-in-the-loop is essential to pre-review customer-facing output.

>>isp+u4
Not really, you can fine tune an LLM to disregard meta instructions / stick to the "core focus" of the chat.

May be a case of moving goalposts, but I'm happy to bet that the speed of movement will slow down to a halt over time.