Posts

Comments

Comment by brockmanmatt on Sufficiently Advanced Language Models Can Do Reinforcement Learning · 2020-08-02T19:06:28.266Z · LW · GW

Ah, sorry, I forgot to add a link to how to evolve the labels. There's a couple different methods in http://gptprompts.wikidot.com/context-stuffing if that helps.

Comment by brockmanmatt on $1000 bounty for OpenAI to show whether GPT3 was "deliberately" pretending to be stupider than it is · 2020-07-22T08:06:26.839Z · LW · GW

I don't think it's a BPE issue but not sure. I'd guess it's closer to the parity issue. It has a hard time implicitly counting in general.

edit: thanks, i know how to link now.

Comment by brockmanmatt on $1000 bounty for OpenAI to show whether GPT3 was "deliberately" pretending to be stupider than it is · 2020-07-21T23:12:45.730Z · LW · GW

It seems to just do really bad with parentheses on their own. It can fix them with like... f(f(f(x))) but not '((())' type situations (I'm just using the beta).

Code: https://gist.github.com/brockmanmatt/aea4fc4a962188f85d83db761bf0ac50