Response Clue Reasoning on Chinese e-commerce user manual dataset
0.544BLEU-1CARE
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CARE2024.08 | 0.544 | 0.45 | 0.387 | 0.339 | 0.497 | 0.484 | |
| Fine-tuned PERTFine-tuning=true2024.08 | 0.433 | 0.345 | 0.292 | 0.254 | 0.438 | 0.396 | |
| ChatGPTFine-tuning=false2024.08 | 0.418 | 0.322 | 0.266 | 0.229 | 0.418 | 0.4 | |
| DRGN2024.08 | 0.405 | 0.312 | 0.247 | 0.206 | 0.385 | 0.356 | |
| TARA2024.08 | 0.397 | 0.301 | 0.239 | 0.195 | 0.361 | 0.327 | |
| PERTFine-tuning=false2024.08 | 0.298 | 0.207 | 0.162 | 0.135 | 0.307 | 0.251 |