Response Appropriateness on Expresso
77Response AppropriatenessPALLM
Evaluation Results
| Method | Links | |
|---|---|---|
| PALLMmode=CLS + GEN2026.03 | 77 | |
| PALLMmode=GEN ONLY2026.03 | 73 | |
| GPT4o-Audio2026.03 | 67.4 | |
| Gemini-2.5 Pro2026.03 | 66.1 | |
| SFTmode=CLS + GEN2026.03 | 65 | |
| SFTmode=GEN ONLY2026.03 | 61 | |
| Qwen-2.52026.03 | 59.6 | |
| Gemma-3n2026.03 | 59 | |
| Gemini-2.5 Flash2026.03 | 51.6 |