List your company

Regression on SDGBiasBench

8.93MAECADE (LLAVA-ONEVISION)
8.326812.398416.4720.5416May 21, 2026

Evaluation results

MethodLinks
CADE (LLAVA-ONEVISION)Params=7B, Setting=CADE2026.058.93-10.792525
CADE (LLAVA-v1.5)Params=13B, Setting=CADE2026.059.16-10.554316.22
CADE (INSTRUCTBLIP)Params=13B, Setting=CADE2026.059.71103712.12
CADE (INTERNVL3)Params=2B, Setting=CADE2026.059.81-5.033727.59
CADE (LLAVA-v1.5)Params=7B, Setting=CADE2026.059.88-5.36356.06
LLAVA-ONEVISIONParams=7B, Setting=Original2026.0510.01—20—
LLAVA-v1.5Params=13B, Setting=Original2026.0510.24—37—
INTERNVL3Params=2B, Setting=Original2026.0510.33—29—
LLAVA-v1.5Params=7B, Setting=Original2026.0510.44—33—
CADE (QWEN2.5-VL)Params=7B, Setting=CADE2026.0510.45-16.273140.91
INSTRUCTBLIPParams=13B, Setting=Original2026.0510.8—33—
GPT-4OSetting=API2026.0511.25—25—
CADE (INSTRUCTBLIP)Params=7B, Setting=CADE2026.0511.45-49.312081.82
CLAUDE 3.5 SONNETSetting=API2026.0511.46—15—
CADE (INTERNVL3)Params=14B, Setting=CADE2026.0511.6-51.692531.58
CADE (QWEN2.5-VL)Params=3B, Setting=CADE2026.0511.93-1.32283.7
QWEN2.5-VLParams=3B, Setting=Original2026.0512.09—27—
GEMINI 2.0 FLASHSetting=API2026.0512.35—12—
QWEN2.5-VLParams=7B, Setting=Original2026.0512.48—22—
CADE (INTERNVL3)Params=8B, Setting=CADE2026.0514.42-27.462042.86
INTERNVL3Params=8B, Setting=Original2026.0519.88—14—
INSTRUCTBLIPParams=7B, Setting=Original2026.0522.59—11—
INTERNVL3Params=14B, Setting=Original2026.0524.01—19—