News for hackers. Stuff that matters.

A comment on SafeGPT: New tool to detect LLMs' hallucinations, biases and privacy issues. Parent comment

Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by MuffinFlavored on Friday April 21, 2023 @10:43AM (#35654513)
> don’t you think OpenAI would’ve worked on something like this?

Along this line of thought: was it a massive oversight for them to not train the model to say "math detected, let me pass that to a solver" instead of trying to guess what token should come next in a math problem?

[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by TazeTSchnitzel on Friday April 21, 2023 @10:46AM (#35654551)
There's a million categories of problem you could ask an LLM to try to solve. You'd need a million solvers…
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by gurleen_s on Friday April 21, 2023 @11:14AM (#35655005)
This seems like a pretty good thing. The model’s ability to detect _which_ solver to use is the killer feature.
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by underlines on Friday April 21, 2023 @11:43AM (#35655492)
you mean huggingGPT?
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by iudqnolq on Friday April 21, 2023 @04:53PM (#35659947)
Why is that a killer feature? Humans are quite good at asking different people different questions. If I need to do a simple math problem I'll just prefix "calculate" and pop it into Google, whereas if I want an intro to a named thing I'll prefix "wikipedia". That's not hard.

GPT is quite useful, but not because it solves the problem of "I don't know where the question I have is answerable by a calculator"

[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by wds on Friday April 21, 2023 @11:05AM (#35654887)
If you used some sort of plugin system, you could just make a solver for your specific task and drop it in. Doesn't ChatGPT Plus do this now?
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by Tostino on Friday April 21, 2023 @11:11AM (#35654970)
It's behind a waitlist.
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by flangola7 on Friday April 21, 2023 @10:55AM (#35654709)
OpenAI plugins can connect with a growing number of things. Zapier is one, which is already several thousand functions.
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by samstave on Friday April 21, 2023 @10:59AM (#35654778)
Can you give some example recipes?
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by BoorishBears on Friday April 21, 2023 @02:47PM (#35658231)
Most obvious one is Wolfram, it passes most math to Wolfram
[ Parent | on HN ]
Re:SafeGPT: New tool to detect LLMs' hallucinations, biases ... by kolinko on Friday April 21, 2023 @12:02PM (#35655807)
They are solving that with plugins now.
[ Parent | on HN ]