brockmanmatt

• brockmanmatt 2 Aug 2020 19:06 UTC

  6 points

  on: Suffi­ciently Ad­vanced Lan­guage Models Can Do Re­in­force­ment Learning

  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.

• brockmanmatt 22 Jul 2020 8:06 UTC

  4 points

  in reply to: ESRogs’s comment on: $1000 bounty for OpenAI to show whether GPT3 was “de­liber­ately” pre­tend­ing to be stupi­der than it is

  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.

• brockmanmatt 21 Jul 2020 23:12 UTC

  7 points

  on: $1000 bounty for OpenAI to show whether GPT3 was “de­liber­ately” pre­tend­ing to be stupi­der than it is

  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