Owain_Evans(Owain Evans)

Model Mis-speci­fi­ca­tion and In­verse Re­in­force­ment Learning

Ma­chine Learn­ing Pro­jects on IDA

Neu­ral nets as a model for how hu­mans make and un­der­stand vi­sual art

Up­date on Ought’s ex­per­i­ments on fac­tored eval­u­a­tion of arguments

Quan­tify­ing House­hold Trans­mis­sion of COVID-19

AI Safety Re­search Pro­ject Ideas

How truth­ful is GPT-3? A bench­mark for lan­guage models

Truth­ful AI: Devel­op­ing and gov­ern­ing AI that does not lie

AMA on Truth­ful AI: Owen Cot­ton-Bar­ratt, Owain Evans & co-authors

The Ra­tion­al­ists of the 1950s (and be­fore) also called them­selves “Ra­tion­al­ists”

Lives of the Cam­bridge poly­math geniuses

How do new mod­els from OpenAI, Deep­Mind and An­thropic perform on Truth­fulQA?

Paper: Teach­ing GPT3 to ex­press un­cer­tainty in words

Paper: Fore­cast­ing world events with neu­ral nets

Paper: On mea­sur­ing situ­a­tional aware­ness in LLMs

Paper: Tell, Don’t Show- Declar­a­tive facts in­fluence how LLMs generalize

How do LLMs give truth­ful an­swers? A dis­cus­sion of LLM vs. hu­man rea­son­ing, en­sem­bles & parrots