# Table Question Answering on WTQ Column Masking ROBUT (test)

Source: [Table Question Answering on WTQ Column Masking ROBUT (test)](<https://www.sota2.com/research/sota/table-question-answering-on-wtq-column-masking-robut-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:34:23.840Z

Evaluation of model performance on the test split of the WTQ dataset using the ROBUT Column Masking perturbation strategy for Table Question Answering.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Table Question Answering |
| Dataset (including any stated split) | WTQ Column Masking ROBUT (test) |
| Metric | Accuracy (Pre-Perturbation) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Accuracy (Pre-Perturbation)](<https://www.sota2.com/research/sota/table-question-answering-on-wtq-column-masking-robut-test.md?metric=Accuracy%20(Pre-Perturbation)>): Higher is better; unit: PERCENT.

- [Accuracy (Post-Perturbation)](<https://www.sota2.com/research/sota/table-question-answering-on-wtq-column-masking-robut-test.md?metric=Accuracy%20(Post-Perturbation)>): Higher is better; unit: PERCENT.

- [Robustness Accuracy (R-ACC)](<https://www.sota2.com/research/sota/table-question-answering-on-wtq-column-masking-robut-test.md?metric=Robustness%20Accuracy%20(R-ACC)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Pre-Perturbation) | Result date | Report paper |
| --- | --- | --- | --- |
| OmniTab | 60.4 | Not recorded | 68fa6fdc1e209131cea47b48 |
| TAPEX | 56.7 | Not recorded | 68fa6fdc1e209131cea47b48 |
| TableFormer | 51 | Not recorded | 68fa6fdc1e209131cea47b48 |
| TAPAS | 45.1 | Not recorded | 68fa6fdc1e209131cea47b48 |
| GPT-3 | 43.1 | Not recorded | 68fa6fdc1e209131cea47b48 |

### OmniTab

Accuracy (Pre-Perturbation): 60.4. Result ID: 6990ef833a6c3148490467f6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fdc1e209131cea47b48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cf31e209131cea36c12.md>)

[Implementation](<https://github.com/jzbjyb/OmniTab>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-tuned"}. This record does not assert independent reproduction.

### TAPEX

Accuracy (Pre-Perturbation): 56.7. Result ID: 6990ef833a6c3148490467f5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fdc1e209131cea47b48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfe1e209131cea31836.md>)

[Implementation](<https://github.com/microsoft/Table-Pretraining>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-tuned"}. This record does not assert independent reproduction.

### TableFormer

Accuracy (Pre-Perturbation): 51. Result ID: 6990ef833a6c3148490467f4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fdc1e209131cea47b48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c7d1e209131cea3443e.md>)

[Implementation](<https://github.com/google-research/tapas/blob/master/TABLEFORMER.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-tuned"}. This record does not assert independent reproduction.

### TAPAS

Accuracy (Pre-Perturbation): 45.1. Result ID: 6990ef833a6c3148490467f3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fdc1e209131cea47b48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c521e209131cea333e0.md>)

[Implementation](<https://github.com/google-research/tapas>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-tuned"}. This record does not assert independent reproduction.

### GPT-3

Accuracy (Pre-Perturbation): 43.1. Result ID: 6990ef833a6c3148490467f7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fdc1e209131cea47b48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa680c1e209131cea19d53.md>)

[Implementation](<https://github.com/openai/gpt-3/blob/master/overlap_frequency.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Few-shot"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.