List your company

Confidence Estimation on DocILE (full)

0.928AUCEXTRACTCONF
0.682560.746280.810.87372Jun 23, 2026

Evaluation results

MethodLinks
EXTRACTCONFsetup=full, type=trained CatBoost classifier, threshold=0.5, id=M62026.060.9280.9740.1990.2370.1360.04391.4
EXTRACTCONFsetup=full, calibration=Lasso Calibration, type=trained CatBoost classifier, threshold=0.5, id=M82026.060.9280.9740.0480.0280.0940.04291.9
EXTRACTCONFsetup=full, calibration=Isotonic Calibration, type=trained CatBoost classifier, threshold=0.5, id=M72026.060.9260.9730.0340.0280.0940.04491.6
All except Mappertype=trained CatBoost classifier, threshold=0.5, id=M52026.060.9110.9670.1840.2230.1410.06589
OCR onlytype=trained CatBoost classifier, threshold=0.5, id=M42026.060.8960.9490.1970.2190.1410.07989.6
Logprobs + Entropytype=trained CatBoost classifier, threshold=0.5, id=M12026.060.880.9530.180.2610.1620.07687.6
Agreement onlytype=trained CatBoost classifier, threshold=0.5, id=M32026.060.8280.9020.2070.2830.1860.10287.5
Spatial onlytype=trained CatBoost classifier, threshold=0.5, id=M22026.060.8140.8970.2580.340.220.10386.5
Self-Consistencytype=model-free baseline, repetition=5x, threshold=0.5, id=B32026.060.7440.8520.0750.0720.1830.13886
Logprob Meantype=model-free baseline, threshold=0.5, id=B12026.060.7050.8640.2450.2450.2530.14584.6
Verbalized Conf.type=model-free baseline, threshold=0.5, id=B22026.060.6920.8280.1630.1630.2130.14684.6