Do LLMs Reason, or Do They Just Predict Math Text?daridor.blog 4 pointsbeagle34 months ago2 commentsSaveHideCopy link On HNComments−gus_massa4moThe best AI is predicting math text better than the me artesanily reasoning the solution. Thats good enough for me.−beagle3OP4moA critique of the KisMATH paper. Bottom line: Headline-claim inflation factor: roughly 3x–4x
Comments
The best AI is predicting math text better than the me artesanily reasoning the solution. Thats good enough for me.
A critique of the KisMATH paper. Bottom line: Headline-claim inflation factor: roughly 3x–4x