Next-Utterance Prediction on adapted IEMOCAP Cross-Scenarios Evaluation Protocol (test)
5.439BLEU-4SayNext-Chat
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SayNext-ChatProtocol=Cross-Scenarios Evaluation2026.01 | 5.439 | 22.292 | 57.625 | 30.76 | 0.5477 | 0.5901 | |
| Gemini2.5Protocol=Cross-Scenarios Evaluation2026.01 | 0.914 | 9.126 | 49.57 | 15.28 | 0.437 | 0.5006 | |
| GPT4oProtocol=Cross-Scenarios Evaluation2026.01 | 0.474 | 9.504 | 47.07 | 17.87 | 0.4366 | 0.4973 | |
| InternVL2Protocol=Cross-Scenarios Evaluation2026.01 | 0.438 | 9.602 | 50.16 | 18.04 | 0.5143 | 0.57 | |
| VideoLLaMA3Protocol=Cross-Scenarios Evaluation2026.01 | 0.379 | 9.528 | 49.15 | 16.74 | 0.447 | 0.5126 | |
| Emotion-LLaMAProtocol=Cross-Scenarios Evaluation2026.01 | 0.257 | 10.058 | 44.78 | 17.24 | 0.5253 | 0.6182 | |
| LLaVA-NeXTProtocol=Cross-Scenarios Evaluation2026.01 | 0.164 | 7.047 | 47.79 | 17.24 | 0.3948 | 0.4654 |