3 ms·
> I didn't know there are many techniques to mitigate this A trivial idea - you can use GPT-3 to inject bullshit/hallucinations into real text. Then train the
by visarga 4y ago
> I didn't know there are many techniques to mitigate this
A trivial idea - you can use GPT-3 to inject bullshit/hallucinations into real text. Then train the model to solve the reverse task, of detecting bullshit in input text.
- stavros 4y agoHow is it going to detect whether a given URL is real, though?
- codetrotter 4y agoGive the AI access to cURL and train it on how to interpret responses. What could possibly go wrong :p
- stavros 4y agoI've been thinking of doing that for some stuff! Teach it how to call APIs and see what happens.
- thornewolf 4y agoI did exactly that! See [1] for my implementation of a telegram bot that is able to write and execute python code. [1] https://github.com/thornewolf/gpt-3-execution https://github.com/thornewolf/gpt-3-execution
- fragmede 4y agoJust give it the return code (200 vs 404 vs NXDOMAIN) and avoid creating Skynet.