Spaces:

Ma-Ri-Ba-Ku
/

alto-llm-corrector

Sleeping

App Files Files Community

alto-llm-corrector / scripts /benchmark.py

Commit History

bench: ECE/Brier calibration harness + blocking ceilings on the real corpus

befdc2b
unverified

Claude Claude Fable 5 commited on 9 days ago

Versioned benchmark + ground-truth corpus seed (P4.2)

baf1728
unverified

Claude Claude Fable 5 commited on 12 days ago