Comment by conception 3 months ago https://cognition.com/frontiercodeQuality vs cost - medium is the sweet (perhaps better too!) spot. 2 comments conception Reply fastball 3 months ago That is just a single benchmark tho dhorthy 3 months ago my issue with frontier code is that it uses a model judge for quality whereas slop code bench forces a model to grapple with its own garbage code in order to receive a functionality reward
dhorthy 3 months ago my issue with frontier code is that it uses a model judge for quality whereas slop code bench forces a model to grapple with its own garbage code in order to receive a functionality reward
That is just a single benchmark tho
my issue with frontier code is that it uses a model judge for quality whereas slop code bench forces a model to grapple with its own garbage code in order to receive a functionality reward