Table Question Answering on ROBUT-WTQ Abbreviation Replacement (test)
82.9Accuracy (Pre-perturbation)TAPEX
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 82.9 | 70.7 | 82.5 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 82.1 | 73.2 | 85.8 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 81.7 | 59.5 | 69.9 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 81 | 57.7 | 66.7 | |
| GPT-3Evaluation Protocol=Few-shot, Sample Size=2002023.06 | 78.5 | 75.1 | 92.3 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 54.8 | 52 | 89.5 | |
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 50.4 | 44.3 | 83.7 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 45.3 | 37.1 | 76.9 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 43.4 | 35.1 | 76.1 | |
| GPT-3Evaluation Protocol=Few-shot2023.06 | 41.5 | 39.2 | 93.8 |