List your company
ResearchFormerly WizWand

Hallucination Evaluation on COCO

53CSSample
17.6426.823645.18Mar 5, 2026
Updated 7mo ago

Evaluation results

MethodLinks
SampleLVLM Model=LLaVA-1.5-7B, Decoding=Sample2026.035315.673.465
SampleLVLM Model=LLaVA-1.5-13B, Decoding=Sample2026.035213.774.966
GreedyLVLM Model=LLaVA-1.5-7B, Decoding=Greedy2026.034913.377.960
GreedyLVLM Model=LLaVA-1.5-13B, Decoding=Greedy2026.0347.812.578.960
AdaIATLVLM Model=LLaVA-1.5-7B, Decoding=Sample2026.034212.373.166
SampleLVLM Model=Qwen2.5-VL-7B, Decoding=Sample2026.034211.372.567
AdaIATLVLM Model=LLaVA-1.5-13B, Decoding=Sample2026.0337.810.17567
AdaIATLVLM Model=Qwen2.5-VL-7B, Decoding=Sample2026.0334.6975.168
PAILVLM Model=LLaVA-1.5-13B, Decoding=Greedy2026.0334.2979.356
SampleLVLM Model=Janus-Pro-7B, Decoding=Sample2026.0334973.564
GreedyLVLM Model=Qwen2.5-VL-7B, Decoding=Greedy2026.0333.68.376.565
PAILVLM Model=Qwen2.5-VL-7B, Decoding=Greedy2026.03328.274.764
PAILVLM Model=LLaVA-1.5-7B, Decoding=Greedy2026.0331.87.877.750
HGAILVLM Model=LLaVA-1.5-7B, Decoding=Greedy2026.0331.46.978.350
AdaIATLVLM Model=LLaVA-1.5-7B, Decoding=Greedy2026.0331.48.379.460
IATLVLM Model=LLaVA-1.5-13B, Decoding=Greedy2026.03318.678.661
HGAILVLM Model=Qwen2.5-VL-7B, Decoding=Greedy2026.0330.4876.865
IATLVLM Model=LLaVA-1.5-7B, Decoding=Greedy2026.0329.8976.861
AdaIATLVLM Model=Janus-Pro-7B, Decoding=Sample2026.0329.88.172.665
IATLVLM Model=Qwen2.5-VL-7B, Decoding=Greedy2026.0329.27.776.467
AdaIATLVLM Model=Qwen2.5-VL-7B, Decoding=Greedy2026.0328.46.976.866
HGAILVLM Model=LLaVA-1.5-13B, Decoding=Greedy2026.0326.47.876.955
GreedyLVLM Model=Janus-Pro-7B, Decoding=Greedy2026.0325.86.776.862
AdaIATLVLM Model=LLaVA-1.5-13B, Decoding=Greedy2026.0325.26.179.260
HGAILVLM Model=Janus-Pro-7B, Decoding=Greedy2026.03215.375.962
IATLVLM Model=Janus-Pro-7B, Decoding=Greedy2026.0320.65.375.365
PAILVLM Model=Janus-Pro-7B, Decoding=Greedy2026.0320.45.676.161
AdaIATLVLM Model=Janus-Pro-7B, Decoding=Greedy2026.03194.976.564