Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Object Hallucination on MSCOCO POPE (test)

85.9Accuracy (Random)

SDCD

-2.56770420.39987343.3674566.335027Jan 7, 2026Jan 10, 2026Jan 14, 2026Jan 18, 2026Jan 22, 2026Jan 26, 2026Jan 30, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.01
85.984.5683.4382.3380.8780.12--93.4677.288.1977.283.3677.13
2026.01
85.6383.3885.0782.8983.9381.76--98.972.0797.0572.3394.5772
2026.01
84.8783.3782.4381.279.979.06--92.5275.8787.3475.8782.5275.87
2026.01
82.9380.8781.179.2478.6377.13--92.0172.1387.972.1382.9672.07
2026.01
82.6779.2382.9779.8981.7778.35--98.866.1397.567.6796.466
2026.01
81.7377.9481.2377.3380.8377.08--98.3764.5397.666495.8464.47
2026.01
0.87870.86430.86370.850.8450.83290.86290.8491------
2026.01
0.87530.86450.84210.8350.80880.80690.84210.8355------
2026.01
0.86840.86830.82650.83370.77310.79280.82270.8316------
2026.01
0.84870.83270.82930.81450.81070.79960.82960.8156------
2026.01
0.83490.82280.79980.79340.76030.76260.79830.7929------