List your company

Hallucination assessment on Object-HalBench

2.6Mention Hallucination RateOmniLMM + RLAIF-V BoN
1.5248.78716.0523.313Dec 1, 2023Jan 11, 2024Feb 21, 2024Apr 2, 2024May 13, 2024Jun 23, 2024Aug 3, 2024

Evaluation results

MethodLinks
OmniLMM + RLAIF-V BoNSize=12B, Feedback=self2024.052.64.5——
OmniLMM + RLAIF-VSize=12B, Feedback=self2024.052.94.5——
LLaVA 1.5 + RLAIF-V BoNSize=7B, Feedback=LLaVA-NeXT2024.053.86.8——
MiniCPM-Llama3-V 2.5Size=8.5B2024.085——10.3
LLaVA 1.5 + RLAIF-VSize=7B, Feedback=LLaVA-NeXT2024.055.210.5——
LLaVA-NeXTSize=34B, Feedback=X2024.056.412.6——
DeepSeek-VL-7BSize=7.3B2024.086.5——11.4
GPT-4V (2023.11.06)2024.087.3——13.6
GPT-4VSize=Unknown, Feedback=Unknown2024.057.313.6——
GPT-4V2023.127.3——13.6
RLHF-VSize=13B, Feedback=Human2024.057.512.2——
RLHF-VBackbone=Vicuna v1.0 13B, Visual Module=BEiT-3, Resolution=4482023.127.5——12.2
MiniCPM-V 2.0Size=2.8B2024.087.8——14.5
MiniGeminiSize=34B, Feedback=X2024.05814.5——
DeepSeek-VL-1.3BSize=1.7B2024.089.6——16.7
OmniLMMSize=12B, Feedback=X2024.0510.919.4——
MiniCPM-V 1.0Size=2.8B2024.0811.5——21.6
Yi-VL-6BSize=6.7B2024.0811.7——19.4
CogVLM-ChatSize=17B2024.0812.6——26.4
SilkieSize=10B, Feedback=GPT-4V2024.0513.427.1——
Yi-VL-34BSize=34B2024.0814——20.7
InstructBLIP2023.1214.3——25.9
Less-is-moreSize=7B, Feedback=X2024.0517.840.3——
LLaVA-RLHFSize=13B, Feedback=Human2024.0518.938.1——
LLaVA-RLHF2023.1218.9——38.1
HA-DPOSize=7B, Feedback=Rule2024.0519.939.9——
Qwen-VL-ChatSize=9.6B2024.0820——43.8
Qwen-VL-Chat2023.1220——43.8
AMP-MEGSize=13B, Feedback=Rule2024.0520.631.7——
Qwen-VL-ChatSize=10B, Feedback=X2024.0520.740.4——
OPERASize=7B, Feedback=X2024.0522.345.1——
LRV2023.1222.3——32.3
LLaVA 1.52023.1222.6——46.3
CCA-LLaVASize=7B, Feedback=X2024.0523.846.7——
VCDSize=7B, Feedback=X2024.0524.348.8——
POVIDSize=7B, Feedback=Rule2024.0524.448.1——
Muffin2023.1224.5——50.5
LLaVA 1.5Size=7B, Feedback=X2024.0527.854.5——
LLaVA2023.1229.5——63
LLaVAParameters=13B2024.03—6329.5—
LLaVA1.5Parameters=7B2024.03—45.923.7—
LLaVA1.5Parameters=13B2024.03—45.221.8—
LLaVA1.5-BPOParameters=7B2024.03—31.915.1—
LLaVA1.5-BPOParameters=13B2024.03—27.312.9—