List your company

Discriminative Hallucination Evaluation on AMBER (test)

86.8AccuracyRC-DPO
69.567274.041178.51582.9889May 27, 2026Jun 2, 2026Jun 9, 2026Jun 15, 2026Jun 22, 2026Jun 28, 2026Jul 5, 2026

Evaluation results

MethodLinks
RC-DPOModel=ThinkLite-VL2026.0586.889.5690.0789.81
OPA-DPOModel=ThinkLite-VL2026.0586.7389.190.5489.81
VCDModel=ThinkLite-VL2026.0586.4689.689.4189.5
RLAIF-VModel=ThinkLite-VL2026.0586.4289.1389.9589.53
VanillaModel=ThinkLite-VL2026.0586.3888.9790.0989.52
RC-DPOModel=MM-Eureka2026.0586.3487.6291.8389.67
OPA-DPOModel=MM-Eureka2026.0586.2387.2892.1189.63
VCDModel=MM-Eureka2026.0586.2287.691.6389.57
SIDModel=MM-Eureka2026.0586.0287.6191.1389.33
VanillaModel=MM-Eureka2026.0585.9987.391.6589.41
SIDModel=ThinkLite-VL2026.0585.7889.488.4988.94
RLAIF-VModel=MM-Eureka2026.0585.6286.4492.2189.22
RC-DPOModel=R1-Onevision2026.0585.2887.5290.0588.76
SeeMeModel=LLaVA-NEXT2026.0783.7——87.8
RegularModel=LLaVA-NEXT2026.0783.6——87.7
DCLAModel=LLaVA-NEXT2026.0783.5——87.7
DoLaModel=LLaVA-NEXT2026.0783.3——87.7
SPINModel=LLaVA-NEXT2026.0783.1——87.5
VCDModel=LLaVA-NEXT2026.0782.6——87.1
OPA-DPOModel=R1-Onevision2026.0577.3686.7676.6481.38
SeeMeModel=mPLUG-Owl22026.0776.8——79.7
DCLAModel=mPLUG-Owl22026.0776.5——79.3
SPINModel=mPLUG-Owl22026.0776.4——78.8
RegularModel=mPLUG-Owl22026.0776.3——78.8
DoLaModel=mPLUG-Owl22026.0776.1——79.1
VCDModel=mPLUG-Owl22026.0775.9——78.4
SeeMeModel=INF-MLLM2026.0773.4——75.1
SeeMeModel=LLaVA-1.52026.0772.7——75.9
DCLAModel=INF-MLLM2026.0772.7——75.1
DCLAModel=LLaVA-1.52026.0772.6——75.7
RLAIF-VModel=R1-Onevision2026.0572.486.0968.3276.18
SPINModel=LLaVA-1.52026.0772.3——75.3
DoLaModel=INF-MLLM2026.0772.2——74.8
VCDModel=LLaVA-1.52026.0772——74.9
RegularModel=LLaVA-1.52026.0771.5——74.1
DoLaModel=LLaVA-1.52026.0771.5——74.2
VCDModel=INF-MLLM2026.0771.5——74.4
RegularModel=INF-MLLM2026.0771.4——74.3
SPINModel=INF-MLLM2026.0770.9——74.1
VCDModel=R1-Onevision2026.0570.6986.0365.2174.18
SIDModel=R1-Onevision2026.0570.3985.4665.2674
VanillaModel=R1-Onevision2026.0570.2385.2865.1673.87