# Table-BERT (GNN-TabFact comparison) Benchmark Scores & Performance

> Table-BERT (GNN-TabFact comparison) has published results in 1 original benchmark tables. The evaluated configuration, source metrics, and precision remain visible below. These results do not produce a general model score or rank.

Canonical page: https://benchlm.ai/models/native-tabfact-table-bert-maintainer

Last updated: 2026-10-01

General benchmark catalog last updated: October 1, 2026. This profile’s source review has its own date above.

## Model Details

| Property | Value |
|----------|-------|
| Creator | TabFact authors |
| Source Type | Research system |
| Reasoning Type | Unspecified |
| Context Window | Not established by evaluation |
| Official model card | [Benchmark-owner results and configuration](https://github.com/wenhuchen/GNN-TabFact) |
| Overall Score | Not computed (source protocol results only) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Table-BERT (GNN-TabFact comparison)
- Variant: benchmark-system
- Benchmarks covered: 0 of 645
- Coverage note: Original benchmark result tables appear below; these metrics are separate from weighted benchmark slots.

## Original benchmark results

[All model results (JSON)](/api/data/benchmarks?model=native-tabfact-table-bert-maintainer) · [Numeric metrics (CSV)](/api/data/benchmarks?model=native-tabfact-table-bert-maintainer&format=csv)

### GNN-TabFact maintainer comparison

The original paper reports released-test accuracy, including simple, complex, and small-test subsets. GNN-TabFact supplies a later maintainer result on the released split. CodaLab’s challenge uses a separate hidden test and publishes six usernames with rounded scores, without identifying their models. Those submissions remain visible but cannot be assigned model profiles from the available source.

Maintainer README comparison. The Table-BERT label does not identify which serialization variant; we preserve that source label separately.

TabFact: A Large-scale Dataset for Table-based Fact Verification — Wenhu Chen and colleagues. Numeric results transcribed and reformatted; evaluation notes and model links added by BenchLM. Published numeric precision is retained.

[Published source](https://github.com/wenhuchen/GNN-TabFact) · [Full TabFact results](/benchmarks/tabfact)

| System | Development accuracy (%) | Test accuracy (%) |
| --- | --- | --- |
| Table-BERT | 66.1 | 65.1 |

## Other TabFact authors Models

- [BERT classifier w/o Table (TabFact)](/models/native-tabfact-bert-classifier-w-o-table) - Score: not computed
- [GNN-TabFact (GNN-TabFact comparison)](/models/native-tabfact-gnn-tabfact-maintainer) - Score: not computed
- [LPA-Ranking w/ Discriminator (Caption) (TabFact)](/models/native-tabfact-lpa-ranking-w-discriminator-caption) - Score: not computed
- [LPA-Ranking w/ Discriminator (TabFact)](/models/native-tabfact-lpa-ranking-w-discriminator) - Score: not computed
- [LPA-Voting w/o Discriminator (TabFact)](/models/native-tabfact-lpa-voting-w-o-discriminator) - Score: not computed
- [LPA-Weighted-Voting (TabFact)](/models/native-tabfact-lpa-weighted-voting) - Score: not computed
- [NSM w/ LPA-guided ML + RL (TabFact)](/models/native-tabfact-nsm-w-lpa-guided-ml-rl) - Score: not computed
- [NSM w/ RL (Binary Reward) (TabFact)](/models/native-tabfact-nsm-w-rl-binary-reward) - Score: not computed
- [Table-BERT-Horizontal-F+T-Concatenate (TabFact)](/models/native-tabfact-table-bert-horizontal-f-t-concatenate) - Score: not computed
- [Table-BERT-Horizontal-F+T-Template (TabFact)](/models/native-tabfact-table-bert-horizontal-f-t-template) - Score: not computed
- [Table-BERT-Horizontal-T+F-Template (TabFact)](/models/native-tabfact-table-bert-horizontal-t-f-template) - Score: not computed
- [Table-BERT-Vertical-F+T-Template (TabFact)](/models/native-tabfact-table-bert-vertical-f-t-template) - Score: not computed
- [Table-BERT-Vertical-T+F-Template (TabFact)](/models/native-tabfact-table-bert-vertical-t-f-template) - Score: not computed