Meme Classification on GOAT-C (test)
81.5AccuracyGemini-3-Flash
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-3-FlashMode=zero-shot2026.06 | 81.5 | 80.4 | |
| Intent ProjectionBackbone=InternVL3-8B2026.06 | 79.4 | 78.3 | |
| Claude-Sonnet-4.6Mode=zero-shot2026.06 | 79.3 | 78.2 | |
| Intent ProjectionBackbone=Qwen3-VL-8B2026.06 | 78.1 | 77 | |
| InternVL3-8BSFT=true, Chain=true2026.06 | 76.8 | 75.6 | |
| Qwen3-VL-8BSFT=true, Chain=true2026.06 | 75.4 | 74.3 | |
| Intent ProjectionBackbone=LLaVA-OneVision-7B2026.06 | 74.8 | 73.7 | |
| GPT-5.4-miniMode=zero-shot2026.06 | 74.6 | 73.1 | |
| LLaVA-OneVision-7BSFT=true, Chain=true2026.06 | 72.3 | 71.1 | |
| InternVL3-8BMode=base2026.06 | 70 | 68.9 | |
| Qwen3-VL-8BMode=base2026.06 | 68.7 | 67.5 | |
| LLaVA-OneVision-7BMode=base2026.06 | 65.4 | 64.2 |