On a fresh, non-leaking Gaokao benchmark, LLM accuracy stays nearly flat across question difficulty, diverging from human Rasch curves and suggesting high scores can overstate human-like ability.
Boone and Amity Noltemeyer
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
GAOKAO-Eval: Does high scores truly reflect strong capabilities in LLMs?
On a fresh, non-leaking Gaokao benchmark, LLM accuracy stays nearly flat across question difficulty, diverging from human Rasch curves and suggesting high scores can overstate human-like ability.