List your company

Harmful Meme Detection on MAMI (test)

84.86AccuracyGPT-5-mini
51.621660.250868.8877.5092Jul 9, 2025Sep 1, 2025Oct 25, 2025Dec 18, 2025Feb 10, 2026Apr 5, 2026May 29, 2026

Evaluation results

MethodLinks
GPT-5-miniVariant=zero-shot, Setup=Base2026.0584.8684.86
GPT-5-miniVariant=in-context, Setup=+ ICL2026.0582.282.14
Qwen-3Variant=MAMI-Sft, Setup=+ ICL2026.0581.981.8
GPT-4.1-miniVariant=zero-shot, Setup=Base2026.0581.981.8
GPT-4oSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.078181
PixtralVariant=Base, Setup=+ LSV2026.0580.780.7
Qwen-3Variant=MAMI-Sft, Setup=Base2026.0580.480.19
Qwen-3Variant=MAMI-Sft, Setup=+ PEFT2026.0580.280.2
U-CoT+Variant=pre-trained, Setup=prompting2026.0579.979.89
Qwen-3Variant=FHM-Sft, Setup=+ PEFT2026.0579.679.59
Qwen-3Variant=MAMI-Sft, Setup=+ LSV2026.0579.478.89
Qwen-3Variant=Base, Setup=+ PEFT2026.0579.179.1
Qwen-3Variant=Base, Setup=+ ICL2026.057979
Qwen-3Variant=Base, Setup=Base2026.0578.778.64
PixtralVariant=MAMI-Sft, Setup=+ LSV2026.0578.578.14
GPT-4.1-miniVariant=in-context, Setup=+ ICL2026.0578.377.98
InternVL-3.5Variant=MAMI-Sft, Setup=Base2026.0578.177.69
InternVL-3.5Variant=MAMI-Sft, Setup=+ LSV2026.0578.177.68
Qwen-3Variant=Base, Setup=+ LSV2026.0577.677.49
PixtralVariant=Base, Setup=+ ICL2026.0576.6776.39
Gemini 1.5 FlashSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0776.474.29
PixtralVariant=MAMI-Sft, Setup=+ ICL2026.057675.19
InternVL-3.5Variant=MAMI-Sft, Setup=+ ICL2026.057675.98
Qwen-3Variant=FHM-Sft, Setup=+ ICL2026.0575.675.57
PixtralVariant=FHM-Sft, Setup=+ LSV2026.0575.375.16
InternVL-3.5Variant=Base, Setup=+ PEFT2026.0574.674.47
InternVL-3.5Variant=FHM-Sft, Setup=+ PEFT2026.0574.674.46
InternVL-3.5Variant=MAMI-Sft, Setup=+ PEFT2026.0574.574.35
InternVL-3.5Variant=FHM-Sft, Setup=+ LSV2026.0574.374.04
InternVL-3.5Variant=Base, Setup=Base2026.0574.274.15
Qwen-3Variant=FHM-Sft, Setup=Base2026.0574.174.1
PixtralVariant=MAMI-Sft, Setup=Base2026.057472.87
PixtralVariant=FHM-Sft, Setup=+ ICL2026.0572.9172.49
InternVL-3.5Variant=FHM-Sft, Setup=Base2026.0572.671.88
InternVL-3.5Variant=FHM-Sft, Setup=+ ICL2026.0572.472.16
Qwen-3Variant=FHM-Sft, Setup=+ LSV2026.0572.372.29
PixtralVariant=FHM-Sft, Setup=Base2026.0572.271.16
InternVL-3.5Variant=Base, Setup=+ ICL2026.0571.871.45
LLaVA-1.6-34BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0771.371.28
InternVL-3.5Variant=Base, Setup=+ LSV2026.0569.767.83
MINDBackbone=LLaVA-1.5-13B, Setting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0768.968.84
PixtralVariant=Base, Setup=Base2026.0567.4263.79
PixtralVariant=Base, Setup=+ PEFT2026.0565.1460.57
PixtralVariant=FHM-Sft, Setup=+ PEFT2026.0564.2359.25
PixtralVariant=MAMI-Sft, Setup=+ PEFT2026.0564.0659.04
LLaVA-1.5-13BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0760.155.52
InstructBLIP-13BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.076057.97
MiniGPT-v2-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0757.452.22
OpenFlamingo-9BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0754.749.88
InstructBLIP-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0753.146.93
LLaVA-1.5-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.0752.941.53