Fact Verification on FEVER 1.0 (dev)
89.07Label AccuracyProoFVer
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ProoFVerTraining Mode=Original, Evidence Retrieval=Gold standard2021.08 | 89.07 | — | |
| CorefBERTTraining Mode=Original, Evidence Retrieval=Gold standard2021.08 | 88.26 | — | |
| ProoFVerTraining Mode=Fine-tuned + L2 Regularization, Evidence Retrieval=Gold standard2021.08 | 87.95 | — | |
| ProoFVerTraining Mode=Fine-tuned, Evidence Retrieval=Gold standard2021.08 | 86.41 | — | |
| KGATTraining Mode=Original, Evidence Retrieval=Gold standard2021.08 | 86.02 | — | |
| CorefBERTTraining Mode=Fine-tuned + L2 Regularization, Evidence Retrieval=Gold standard2021.08 | 84.22 | — | |
| KGATTraining Mode=Fine-tuned + L2 Regularization, Evidence Retrieval=Gold standard2021.08 | 79.93 | — | |
| CorefBERTTraining Mode=Fine-tuned, Evidence Retrieval=Gold standard2021.08 | 78.79 | — | |
| KGATEncoder Backbone=ROBERTa Large2019.10 | 78.29 | 76.11 | |
| KGATEncoder Backbone=BERT Base2019.10 | 78.02 | 75.88 | |
| KGATEncoder Backbone=BERT Large2019.10 | 77.91 | 75.86 | |
| KGATTraining Mode=Fine-tuned, Evidence Retrieval=Gold standard2021.08 | 76.67 | — | |
| KGATEncoder Backbone=BERT Base, Retrieval Method=ESIM Retrieval2019.10 | 75.51 | 71.61 | |
| GATEncoder Backbone=BERT Base, Retrieval Method=ESIM Retrieval2019.10 | 75.13 | 71.04 | |
| SR-MRS2019.10 | 75.12 | 70.18 | |
| GEAR2019.10 | 74.84 | 70.69 | |
| BERT (Large)Encoder Backbone=BERT Large2019.10 | 74.59 | 72.42 | |
| BERT Concat2019.10 | 73.67 | 68.89 | |
| BERT (Base)Encoder Backbone=BERT Base2019.10 | 73.51 | 71.38 | |
| BERT Pair2019.10 | 73.3 | 68.9 | |
| UNC NLP2019.10 | 69.72 | 66.49 | |
| UCL MRG2019.10 | 69.66 | 65.41 | |
| Athene2019.10 | 68.49 | 64.74 |