Papers › Can BERT eat RuCoLA? Topological Data Analysis to Explain
Can BERT eat RuCoLA? Topological Data Analysis to Explain
Irina Proskurina, Irina Piontkovskaya, Ekaterina Artemova
This paper investigates how Transformer language models (LMs) fine-tuned for acceptability classification capture linguistic features. Our approach uses the best practices of topological data analysis (TDA) in NLP: we construct directed attention graphs from attention matrices, derive topological features from them, and feed them to linear classifiers. We introduce two novel features, chordality, and the matching number, and show that TDA-based classifiers outperform fine-tuning baselines. We experiment with two datasets, CoLA and RuCoLA in English and Russian, typologically different languages. On top of that, we propose several black-box introspection techniques aimed at detecting changes in the attention mode of the LMs during fine-tuning, defining the LM's prediction confidences, and associating individual heads with fine-grained grammar phenomena. Our results contribute to understanding the behavior of monolingual LMs in the acceptability classification task, provide insights into the functional roles of attention heads, and highlight the advantages of TDA-based approaches for analyzing LMs. We release the code and the experimental results for further uptake.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
1 archive task tag without a task page not shown.
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Linguistic Acceptability | CoLA | BERT+TDA | Accuracy | 88.2% | #2 of 43 | Archive leaderboard | report |
| Linguistic Acceptability | CoLA | BERT+TDA | MCC | 0.726 | #2 of 43 | Archive leaderboard | report |
| Linguistic Acceptability | CoLA | RoBERTa+TDA | Accuracy | 87.3% | #3 of 43 | Archive leaderboard | report |
| Linguistic Acceptability | CoLA | RoBERTa+TDA | MCC | 0.695 | #3 of 43 | Archive leaderboard | report |
| Linguistic Acceptability | RuCoLA | Ru-RoBERTa+TDA | Accuracy | 85.7 | #1 of 9 | Archive leaderboard | report |
| Linguistic Acceptability | RuCoLA | Ru-RoBERTa+TDA | MCC | 0.594 | #1 of 9 | Archive leaderboard | report |
| Linguistic Acceptability | RuCoLA | Ru-BERT+TDA | Accuracy | 80.1 | #3 of 9 | Archive leaderboard | report |
| Linguistic Acceptability | RuCoLA | Ru-BERT+TDA | MCC | 0.478 | #3 of 9 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections