---
title: "OLEA — paper by Grace et al. (2023) — summary and ideas"
description: "Offensive-language classifiers do well when a message has obvious slurs or keywords. 23 references, cited by 1 paper here."
authors: ["Marie Grace", "Jay Seabrum", "Dananjay Srinivas", "Alexis Palmer"]
year: 2023
venue: "Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations"
source: "https://aclanthology.org/2023.eacl-demo.24/"
doi: "10.18653/v1/2023.eacl-demo.24"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/grace-2023-olea-tool-infrastructure-offensive-language/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/grace-2023-olea-tool-infrastructure-offensive-language/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/14/) / OLEA: Tool and Infrastructure for Offensive Language Error Analysis in English

# OLEA: Tool and Infrastructure for Offensive Language Error Analysis in English

**Authors:** [Marie Grace](https://md.research.elamur.ai/free/en/authors/marie-grace/), Jay Seabrum, [Dananjay Srinivas](https://md.research.elamur.ai/free/en/authors/dananjay-srinivas/) and [Alexis Palmer](https://md.research.elamur.ai/free/en/authors/alexis-palmer/) · **Year:** 2023 · **Venue:** [Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations](https://md.research.elamur.ai/free/en/venues/eacl/) ([EACL 2023](https://md.research.elamur.ai/free/en/venues/eacl/2023/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.eacl-demo.24/) · [PDF](https://aclanthology.org/2023.eacl-demo.24.pdf) · [DOI 10.18653/v1/2023.eacl-demo.24](https://doi.org/10.18653/v1/2023.eacl-demo.24) · [OpenAlex](https://openalex.org/W4386566846)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · 1 in our library, of the 149,439 papers whose bibliographies we read

## Summary
Offensive-language classifiers do well when a message has obvious slurs or keywords. They break down on implicit abuse, negation, dialectal text such as African American English, and cases where annotators disagree. A single F1 score hides these weaknesses. The authors built OLEA, an open-source Python library aimed at model developers rather than end users. It lets them break down a classifier's errors by linguistically meaningful categories with very little code.

The user loads predictions into a Pandas dataframe and maps the model's labels to the dataset's labels. OLEA then compares coarse offensive/non-offensive performance across subsets grouped by a feature, such as slur present vs. absent, and plots the results with Matplotlib.

Two diagnostic resources come preloaded through the HuggingFace Hub: HateCheck and COLD. A class-based scaffold lets anyone add new datasets, such as OLID, and pair them with custom analyses.

The contribution is infrastructure rather than a new model. It gives a shared, reproducible way to see what detectors cannot handle, and the authors hope it grows into a community library for error analysis.

## Key points
- OLEA is published on PyPI as `olea`, with code on GitHub under the MIT License. It is built on Pandas, Matplotlib and HuggingFace `datasets`.
- It includes two preloaded datasets:
  - **HateCheck**: 3,901 test cases generated from 866 templates, covering seven protected groups, with labels for functionality and target identity.
  - **COLD**: binary labels for offensiveness, slur, adjectival nominalization and linguistic distancing. Each label comes from three annotators plus a majority vote.
- It offers five built-in analyses:
  - `analyze_on`: performance by any categorical column, with versions tuned for COLD and HateCheck.
  - `check_anno_agreement`: compares items where all annotators agreed with items where only some did.
  - `aave`: groups items using TwitterAAE dialect scores.
  - `check_substring`: compares items with and without a given string.
  - `str_len_analysis`: a histogram of performance by text length.
- Extension works through a `Dataset` subclass (demonstrated with OLID level-A) and an optional custom Analysis class. Submissions are sanity-checked and wrapped in a `DatasetSubmissionObject`.
- Stated limitations:
  - Substring analysis assumes words are separated by spaces.
  - The tool focuses on the offensive-language domain.
  - It does not correct annotator or sociolinguistic biases in the data.
  - Offensive examples are hidden by default (`show_examples=False`).

## Interesting ideas
- Break one aggregate score into scores per linguistic phenomenon, using annotated feature columns, so weak spots become visible instead of averaged away
- Treat full annotator agreement as a sign of an easy item and partial agreement as a sign of a hard one, then compare performance on the two groups
- Compare performance on inputs that contain a chosen substring with inputs that do not, as a cheap targeted probe
- Slice evaluation by an automatically inferred dialect score to expose bias toward standard varieties
- Combine a per-dataset adapter with an explicit label-mapping step and sanity checks on submitted outputs, so new evaluation sets plug in with minimal code

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): report parser and frame accuracy separately for negation, passives, relative clauses, coordination and similar phenomena, so regressions in rare constructions are not hidden by overall LAS.
- **I2** (learn): split silver data by whether all witness parsers agreed or only a majority did, and track accuracy on the contested subset to judge how trustworthy the silver labels are.
- **I3** (tokens): check tagger and lemma accuracy on sentences containing ambiguous words like "that" or "'s" against sentences without them, to find token-level failure points quickly.
- **I4** (tree): score parsing quality on informal or dialectal English separately, to see whether the treebank-trained parser degrades on non-standard text.
- **I5** (tree): use the same pattern for treebank loaders and UD dialect conversion, declaring label mappings per corpus and validating converted trees before they reach training.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of Colorado Boulder](https://md.research.elamur.ai/free/en/universities/university-of-colorado-boulder/) — University, Boulder, United States · 69 works · 51,069 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,772 works · [🤝 Communication](https://md.research.elamur.ai/free/en/shelves/social-sciences/communication/) · 97 works

🧭 Topics (2): [English](https://md.research.elamur.ai/free/en/topics/english/) · 2,263 works · [😊 Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · 972 works

## Cite

Cite as: Marie Grace, Jay Seabrum, Dananjay Srinivas and Alexis Palmer. 2023. OLEA: Tool and Infrastructure for Offensive Language Error Analysis in English. Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{grace2023olea,
  title = {{OLEA: Tool and Infrastructure for Offensive Language Error Analysis in English}},
  author = {Marie Grace, Jay Seabrum, Dananjay Srinivas and Alexis Palmer},
  booktitle = {Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations},
  year = {2023},
  doi = {10.18653/v1/2023.eacl-demo.24},
  url = {https://doi.org/10.18653/v1/2023.eacl-demo.24}
}
```

- [← previous Noor-Ghateh: A Benchmark Dataset for Evaluating Arabic Word Segmenters in Hadith Domain](https://md.research.elamur.ai/free/en/papers/alshuhayeb-2023-noor-ghateh-benchmark-dataset-evaluating/)
- [next → Overview of the HASOC Subtrack at FIRE 2023: Identification of Tokens Contributing to Explicit Hate in English by Span Detection](https://md.research.elamur.ai/free/en/papers/masud-2023-overview-hasoc-subtrack-fire-2023/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/14/)

## References

23 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Alexis Palmer, Christine Carr, Melissa Robinson, and Jordan Sanders. 2020. COLD: Annotation scheme and evaluation data set for complex offensive language in English. Journal for Language Technology and Computational Linguistics, 34(1):1–28. [↗](https://doi.org/10.21248/jlcl.34.2020.222)
2. Francesco Barbieri, José Camacho-Collados, Leonardo Neves, and Luis Espinosa Anke. 2020. [Tweeteval: Unified benchmark and comparative evaluation for tweet classification](https://md.research.elamur.ai/free/en/papers/barbieri-2020-tweeteval-unified-benchmark-comparative-evaluation/). CoRR, abs/2010.12421. [↗](https://arxiv.org/abs/2010.12421)
3. Su Lin Blodgett, Lisa Green, and Brendan O’Connor. 2016. [Demographic dialectal variation in social media: A case study of African-American English](https://md.research.elamur.ai/free/en/papers/blodgett-2016-demographic-dialectal-variation-social-media/). In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1119–1130, Austin, Texas. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/d16-1120)
4. F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay. 2011. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830. [↗](https://doi.org/10.48550/arxiv.1201.0490)
5. Tommaso Caselli, Valerio Basile, Jelena Mitrović, Inga Kartoziya, and Michael Granitzer. 2020. [I feel offended, don’t be abusive!](https://md.research.elamur.ai/free/en/papers/caselli-2020-i-feel-offended-don-t/) implicit/explicit messages in offensive and abusive language. In Proceedings of the 12th Language Resources and Evaluation Conference, pages 6193–6202, Marseille, France. European Language Resources Association.
6. Paul Röttger, Bertram Vidgen, Dong Nguyen, Zeerak Waseem, Helen Margetts, and Janet B Pierrehumbert. 2020. [Hatecheck: Functional tests for hate speech detection models](https://md.research.elamur.ai/free/en/papers/rottger-2021-hatecheck-functional-tests-hate-speech/). arXiv preprint arXiv:2012.15606. [↗](https://arxiv.org/abs/2012.15606)
7. Mai ElSherief, Caleb Ziems, David Muchlinski, Vaishnavi Anupindi, Jordyn Seybolt, Munmun De Choudhury, and Diyi Yang. 2021. [Latent hatred: A benchmark for understanding implicit hate speech](https://md.research.elamur.ai/free/en/papers/elsherief-2021-latent-hatred-benchmark-understanding-implicit/). In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 345–363, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. [↗](https://doi.org/10.48550/arxiv.2109.05322)
8. Komal Florio, Valerio Basile, Marco Polignano, Pierpaolo Basile, and Viviana Patti. 2020. Time of your hate: The challenge of time in hate speech detection on social media. Applied Sciences, 10(12):4180. [↗](https://doi.org/10.3390/app10124180)
9. Eduard Hovy. 2022. Rediscovering the need for representation and knowledge. ACL 2022 - 60th Annual Meeting of the Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Eduard+Hovy%2E+2022%2E+Rediscovering+the+need+for+representation+and+knowledge%2E+ACL+2022+%2D+60th+Annual+Meeting+of+the+Association+for+Computational+Linguistics%2E)
10. J. D. Hunter. 2007. Matplotlib: A 2d graphics environment. Computing in Science & Engineering, 9(3):90–95. [↗](https://doi.org/10.1109/mcse.2007.55)
11. Chris J Kennedy, Geoff Bacon, Alexander Sahn, and Claudia von Vacano. 2020. Constructing interval variables via faceted rasch measurement and multi-task deep learning: a hate speech application. arXiv preprint arXiv:2009.10277. [↗](https://arxiv.org/abs/2009.10277)
12. Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, and Animesh Mukherjee. 2021. Hatexplain: A benchmark dataset for explainable hate speech detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 14867–14875. [find](https://scholar.google.com/scholar?q=Binny+Mathew%2C+Punyajoy+Saha%2C+Seid+Muhie+Yimam%2C+Chris+Biemann%2C+Pawan+Goyal%2C+and+Animesh+Mukherjee%2E+2021%2E+Hatexplain%3A+A+benchmark+dataset+for+explainable+hate+speech+detection%2E+In+Proceedings+of+the+AAA)
13. Angelina McMillan-Major, Amandalynne Paullada, and Yacine Jernite. 2022. An interactive exploratory tool for the task of hate speech detection. In Proceedings of the Second Workshop on Bridging Human– Computer Interaction and Natural Language Processing, pages 11–20, Seattle, Washington. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Angelina+McMillan%2DMajor%2C+Amandalynne+Paullada%2C+and+Yacine+Jernite%2E+2022%2E+An+interactive+exploratory+tool+for+the+task+of+hate+speech+detection%2E+In+Proceedings+of+the+Second+Workshop+on+Bridging+Human%E2%80%93)
14. Maarten Sap, Swabha Swayamdipta, Laura Vianna, Xuhui Zhou, Yejin Choi, and Noah A Smith. 2021. [Annotators with attitudes: How annotator beliefs and identities bias toxic language detection](https://md.research.elamur.ai/free/en/papers/sap-2022-annotators-attitudes-how-annotator-beliefs/). arXiv preprint arXiv:2111.07997. [↗](https://arxiv.org/abs/2111.07997)
15. Anna Schmidt and Michael Wiegand. 2017. [A survey on hate speech detection using natural language processing](https://md.research.elamur.ai/free/en/papers/schmidt-2017-survey-hate-speech-detection-natural/). In Proceedings of the Fifth International Workshop on Natural Language Processing for Social Media, pages 1–10, Valencia, Spain. Association for Computational Linguistics.
16. Bertie Vidgen, Dong Nguyen, Helen Margetts, Patricia Rossini, and Rebekah Tromble. 2021. [Introducing CAD: the contextual abuse dataset](https://md.research.elamur.ai/free/en/papers/vidgen-2021-introducing-cad-contextual-abuse-dataset/). In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2289–2303, Online. Association for Computational Linguistics.
17. Zeerak Waseem, Thomas Davidson, Dana Warmsley, and Ingmar Weber. 2017a. [Understanding abuse: A typology of abusive language detection subtasks](https://md.research.elamur.ai/free/en/papers/waseem-2017-understanding-abuse-typology-abusive-language/). arXiv preprint arXiv:1705.09899. [↗](https://arxiv.org/abs/1705.09899)
18. Zeerak Waseem, Thomas Davidson, Dana Warmsley, and Ingmar Weber. 2017b. [Understanding abuse: A typology of abusive language detection subtasks](https://md.research.elamur.ai/free/en/papers/waseem-2017-understanding-abuse-typology-abusive-language/). In Proceedings of the First Workshop on Abusive Language Online, pages 78–84, Vancouver, BC, Canada. Association for Computational Linguistics.
19. Wes McKinney. 2010. Data Structures for Statistical Computing in Python. In Proceedings of the 9th Python in Science Conference, pages 56 – 61. [find](https://scholar.google.com/scholar?q=Wes+McKinney%2E+2010%2E+Data+Structures+for+Statistical+Computing+in+Python%2E+In+Proceedings+of+the+9th+Python+in+Science+Conference%2C+pages+56+%E2%80%93+61%2E)
20. Michael Wiegand, Josef Ruppenhofer, and Thomas Kleinbauer. 2019. [Detection of Abusive Language: the Problem of Biased Datasets](https://md.research.elamur.ai/free/en/papers/wiegand-2019-detection-abusive-language-problem-biased/). In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 602–608, Minneapolis, Minnesota. Association for Computational Linguistics.
21. Wenjie Yin and Arkaitz Zubiaga. 2021. Towards generalisable hate speech detection: a review on obstacles and solutions. PeerJ Computer Science, 7:e598. [↗](https://doi.org/10.7717/peerj-cs.598)
22. Marcos Zampieri, Shervin Malmasi, Preslav Nakov, Sara Rosenthal, Noura Farra, and Ritesh Kumar. 2019a. [Predicting the Type and Target of Offensive Posts in Social Media](https://md.research.elamur.ai/free/en/papers/zampieri-2019-predicting-type-target-offensive-posts/). In Proceedings of NAACL.
23. Marcos Zampieri, Shervin Malmasi, Preslav Nakov, Sara Rosenthal, Noura Farra, and Ritesh Kumar. 2019b. [SemEval-2019 task 6: Identifying and categorizing offensive language in social media (OffensEval)](https://md.research.elamur.ai/free/en/papers/zampieri-2019-semeval-2019-task-6-identifying/). In Proceedings of the 13th International Workshop on Semantic Evaluation, pages 75–86, Minneapolis, Minnesota, USA. Association for Computational Linguistics.

*Updated 1 November 2026*