Harmful Meme Detection on HarM
83.82AccuracyMOMENTA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MOMENTATraining/Evaluation Protocol=Training-based2025.07 | 83.82 | 82.8 | |
| MOMENTAProtocol=Supervised / Training-Based2026.05 | 83.32 | 82.8 | |
| Late FusionProtocol=Supervised / Training-Based2026.05 | 73.24 | 70.25 | |
| Late FusionTraining/Evaluation Protocol=Training-based2025.07 | 73.21 | 70.25 | |
| PrismAgentProtocol=Zero-Shot, Model Category=Agent-Based Methods (Open-Source), Backbone=LLaVA-1.6-34B2026.05 | 71.19 | 69.78 | |
| PrismAgentProtocol=Zero-Shot, Model Category=Agent-Based Methods (Open-Source), Backbone=LLaVA-1.5-13B2026.05 | 70.62 | 68.44 | |
| MINDBackbone=LLaVA-1.5-13B, Training/Evaluation Protocol=Training-free2025.07 | 68.93 | 65.19 | |
| MINDBackbone=LLaVA-1.5-13B, Setting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 68.93 | 65.19 | |
| MINDProtocol=Zero-Shot, Model Category=Agent-Based Methods (Open-Source), Backbone=LLaVA-1.5-13B2026.05 | 68.93 | 65.19 | |
| GPT-4oSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 67.51 | 60.29 | |
| LLaVA-1.6-34BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 67.51 | 61.59 | |
| GPT-4oProtocol=Zero-Shot, Model Category=Closed-Source VLMs2026.05 | 67.51 | 60.29 | |
| LLaVA-1.6-34BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 67.51 | 61.59 | |
| Gemini 1.5 FlashSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 66.1 | 64.18 | |
| InstructBLIP-13BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 64.92 | 49.61 | |
| InstructBLIP-13BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 64.92 | 49.61 | |
| Gemini-2.0-FlashProtocol=Zero-Shot, Model Category=Closed-Source VLMs2026.05 | 64.67 | 58.84 | |
| OpenFlamingo-9BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 63.42 | 54.36 | |
| OpenFlamingo-9BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 63.42 | 54.36 | |
| LLaVA-1.5-13BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 62.28 | 50.45 | |
| LLaVA-1.5-13BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 62.28 | 50.45 | |
| MiniGPT-v2-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 60.12 | 52.39 | |
| MiniGPT-V-2BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 60.12 | 52.39 | |
| LLaVA-1.5-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 59.23 | 49.44 | |
| LLaVA-1.5-7BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 59.23 | 49.44 | |
| InstructBLIP-7BProtocol=Zero-Shot, Model Category=Open-Source VLMs2026.05 | 51.53 | 50.99 | |
| InstructBLIP-7BSetting=Zero-shot, Prompt Strategy=Chain-of-Thought2025.07 | 51.13 | 50.99 |