Factual Consistency Evaluation on QAGS-CNN (test)
0.735Pearson rBARTScore
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BARTScore2026.06 | 0.735 | 0.68 | 0.557 | |
| UniEvalBackbone=T52026.06 | 0.682 | 0.662 | 0.532 | |
| BINEVALBackbone=Claude2026.06 | 0.665 | 0.702 | 0.597 | |
| BINEVALBackbone=gpt-oss2026.06 | 0.651 | 0.642 | 0.551 | |
| G-EvalBackbone=GPT-42026.06 | 0.631 | 0.685 | 0.591 | |
| CTC2026.06 | 0.619 | 0.564 | 0.45 | |
| BERTScore2026.06 | 0.576 | 0.505 | 0.399 | |
| QAGS2026.06 | 0.545 | — | — | |
| FactCC2026.06 | 0.416 | 0.484 | 0.376 | |
| UniEvalBackbone=gpt-oss2026.06 | 0.415 | 0.382 | 0.357 | |
| MoverScore2026.06 | 0.414 | 0.347 | 0.271 | |
| ROUGE-L2026.06 | 0.357 | 0.324 | 0.254 | |
| G-EvalBackbone=gpt-oss2026.06 | 0.045 | 0.028 | 0.027 |