Harmful Meme Detection on ToxiCN
79.21AccuracyALARM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ALARMTraining Free=true, Setting=Label-Free, Backbone=Qwen2.5-VL-72B2025.12 | 79.21 | 72.51 | |
| MOMENTATraining Free=false, Setting=Label-Driven2025.12 | 77.87 | 70.25 | |
| ALARMTraining Free=true, Setting=Label-Free, Backbone=GPT-4o2025.12 | 77.45 | 67.87 | |
| MR.HARMTraining Free=false, Setting=Label-Driven2025.12 | 77.16 | 69.55 | |
| PromptHateTraining Free=false, Setting=Label-Driven2025.12 | 76.04 | 67.45 | |
| Pro-CapTraining Free=false, Setting=Label-Driven2025.12 | 75.7 | 71.36 | |
| LoReHMTraining Free=true, Setting=Few Shot, Backbone=Qwen2.5-VL-72B2025.12 | 75.29 | 64.36 | |
| ExplainHMTraining Free=false, Setting=Label-Driven2025.12 | 75.2 | 67.6 | |
| ISMTraining Free=false, Setting=Label-Driven2025.12 | 74.92 | 67.02 | |
| Qwen2.5-VL-72BTraining Free=true, Setting=Few Shot2025.12 | 74.08 | 66.4 | |
| LoReHMTraining Free=true, Setting=Few Shot, Backbone=GPT-4o2025.12 | 72.55 | 57.85 | |
| HHPromptTraining Free=false, Setting=Label-Driven2025.12 | 72.29 | 60.03 | |
| GPT-4oTraining Free=true, Setting=Few Shot2025.12 | 71.26 | 50.19 | |
| Mod-HATETraining Free=false, Setting=Few Shot2025.12 | 61.24 | 59.71 | |
| OPT-30BTraining Free=true, Setting=Few Shot2025.12 | 49.75 | 48.99 | |
| OpenFlamingo-9BTraining Free=true, Setting=Few Shot2025.12 | 49.01 | 48.69 |