Complex Emotion Interpretation on EIBench
39.27Recall (LLaMA-3 Criteria)ChatGPT-4o
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ChatGPT-4oSource=Close-Source2025.04 | 39.27 | 39.57 | |
| LLaVA-NEXTSource=Open-Source, Model Scale=13B2025.04 | 39.16 | 33.6 | |
| LLaVA-NEXTSource=Open-Source, Model Scale=7B2025.04 | 38.71 | 33.5 | |
| LLaVA-1.5Source=Open-Source, Model Scale=13B2025.04 | 38.1 | 39.53 | |
| LLaVA-NEXTSource=Open-Source, Model Scale=34B2025.04 | 35.37 | 33.1 | |
| MiniGPT-v2Source=Open-Source2025.04 | 35.1 | 36 | |
| Video-LLaVASource=Open-Source2025.04 | 30.9 | 32.27 | |
| ChatGPT-4VSource=Close-Source2025.04 | 28 | 30.6 | |
| OtterSource=Open-Source2025.04 | 27.9 | 33.23 | |
| Claude-3 haikuSource=Close-Source2025.04 | 24 | 24.77 | |
| Qwen-VL-ChatSource=Open-Source2025.04 | 22 | 32.4 | |
| Claude-3 sonnetSource=Close-Source2025.04 | 21.37 | 22.4 | |
| Qwen-vl-plusSource=Close-Source2025.04 | 20.37 | 19.6 |