Half of the benchmarks don't have published answers. That's why training data pr... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		fsh 6 days ago \| parent \| context \| favorite \| on: How good are frontier models at physics? Half of the benchmarks don't have published answers. That's why training data providers have been hiring physicists to solve them.
		help

red75prime 5 days ago | [–]

This is unconventional benchmaxxing then, when they decrease the benchmark scores to allow models to generalize on correct solutions.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact