---
title: "GKnow — paper by Veloso and Schuetze (2026)"
description: "The authors build GKnow, an English benchmark for autoregressive models. 47 references."
authors: ["Leonor Veloso", "Hinrich Schuetze"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.466/"
doi: "10.18653/v1/2026.acl-long.466"
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/veloso-2026-gknow-measuring-entanglement-gender-bias/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/veloso-2026-gknow-measuring-entanglement-gender-bias/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/31/) / GKnow: Measuring the Entanglement of Gender Bias and Factual Gender

# GKnow: Measuring the Entanglement of Gender Bias and Factual Gender

**Authors:** Leonor Veloso and [Hinrich Schuetze](https://md.research.elamur.ai/free/en/authors/hinrich-schuetze/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **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/2026.acl-long.466/) · [PDF](https://aclanthology.org/2026.acl-long.466.pdf) · [DOI 10.18653/v1/2026.acl-long.466](https://doi.org/10.18653/v1/2026.acl-long.466) · [OpenAlex](https://openalex.org/W7166852578)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Work on the inner mechanics of gender in language models usually studies one narrow task, such as pronoun prediction. It also tends to blur two things: gender that a word actually encodes (*sister* → *she*) and gender inferred from a stereotype (*nurse* → *she*). The authors build GKnow, an English benchmark for autoregressive models. Each prompt pairs a type of gendered subject (pronoun, gender label, name, factually gendered noun, or stereotypical occupation/trait) with a type of gendered output. Every prompt is labelled stereotypical or factual.

Using EAP-IG circuit discovery on Llama-3.1-8b and Olmo-7b, they find that stereotypical and factual gender circuits overlap heavily. Each circuit can largely perform the other's task. They then ablate top neurons selected with integrated gradients and evaluate on GKnow, StereoSet and DiFair. The ablation reliably damages factual gender predictions. Yet on bias benchmarks it can look like successful debiasing, because anti-stereotypical completions become more likely. The authors therefore argue that neuron ablation alone is an unreliable debiasing method. They also argue that bias benchmarks should account for factual gender.

## Key points
- **GKnow size:** the full benchmark has 91,490 examples; the experiments use a reduced version of 6992. Data and code are at github.com/leonorv/gknow.
- **Circuit analysis:** EAP-IG circuits show high overlap and high cross-task faithfulness between stereotypical and factual gender. Similarity is lower across different task types, such as pronoun prediction versus gender prediction.
- **Ablation results:** removing the top 10 or top 50 IG neurons lowers the probability of the correct factual completion. For Olmo, the top-50 ablation reduced it on GKnow Factual from about 90.1 to 82.8. On StereoSet and GKnow Stereo, the probability of the opposite-gender completion rises, which can mask this damage.
- **Interpretable neurons:** some neurons are shared across several GKnow subsets. Examples include Olmo L31N8077 (masculine tokens versus *she*/*her*) and a Llama neuron promoting male names.
- **Limitations:** the study covers English only, binary gender, occupational and trait stereotypes, and relatively small models.

*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

- [Ludwig-Maximilians-Universität München](https://md.research.elamur.ai/free/en/universities/ludwig-maximilians-universitat-munchen/) — University, Munich, Germany · 107 works · 56,478 citations

📚 Shelves (6): [🖥️ 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 · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,772 works · [🤝 Safety Research](https://md.research.elamur.ai/free/en/shelves/social-sciences/safety-research/) · 543 works

🧭 Topics (1): [🛡️ Bias, fairness & safety](https://md.research.elamur.ai/free/en/topics/bias-fairness-safety/) · 517 works

## Cite

Cite as: Leonor Veloso and Hinrich Schuetze. 2026. GKnow: Measuring the Entanglement of Gender Bias and Factual Gender. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

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

```bibtex
@inproceedings{veloso2026gknow,
  title = {{GKnow: Measuring the Entanglement of Gender Bias and Factual Gender}},
  author = {Leonor Veloso and Hinrich Schuetze},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.466},
  url = {https://doi.org/10.18653/v1/2026.acl-long.466}
}
```

- [← previous Giving Voice to the Constitution: Low-Resource Text-to-Speech for Quechua and Spanish Using a Bilingual Legal Corpus](https://md.research.elamur.ai/free/en/papers/ortega-2026-giving-voice-constitution-low-resource/)
- [next → Gladiators at #SMM4H–HeaRD 2026: Multi-Seed XLM-RoBERTa Ensemble with Focal Loss and Per-Language Threshold Optimization for Multilingual Adverse Drug Event Detection](https://md.research.elamur.ai/free/en/papers/singh-2026-gladiators-smm4h-heard-2026-multi/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/31/)

## References

47 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Marion Bartl and Susan Leavy. 2022. Inferring gender: A scalable methodology for gender detection with online lexical databases. In Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion, pages 47–58. [find](https://scholar.google.com/scholar?q=Marion+Bartl+and+Susan+Leavy%2E+2022%2E+Inferring+gender%3A+A+scalable+methodology+for+gender+detection+with+online+lexical+databases%2E+In+Proceedings+of+the+Second+Workshop+on+Language+Technology+for+Equali)
2. Nora Belrose, David Schneider-Joseph, Shauli Ravfogel, Ryan Cotterell, Edward Raff, and Stella Biderman. 2024. Leace: Perfect linear concept erasure in closed form. Advances in Neural Information Processing Systems, 36. [find](https://scholar.google.com/scholar?q=Nora+Belrose%2C+David+Schneider%2DJoseph%2C+Shauli+Ravfogel%2C+Ryan+Cotterell%2C+Edward+Raff%2C+and+Stella+Biderman%2E+2024%2E+Leace%3A+Perfect+linear+concept+erasure+in+closed+form%2E+Advances+in+Neural+Information+Proc)
3. Talia Mae Bettcher. 2013. Trans women and the meaning of “woman”. [find](https://scholar.google.com/scholar?q=Talia+Mae+Bettcher%2E+2013%2E+Trans+women+and+the+meaning+of+%E2%80%9Cwoman%E2%80%9D%2E)
4. Su Lin Blodgett, Gilsinia Lopez, Alexandra Olteanu, Robert Sim, and Hanna Wallach. 2021. Stereotyping norwegian salmon: An inventory of pitfalls in fairness benchmark datasets. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint… [find](https://scholar.google.com/scholar?q=Su+Lin+Blodgett%2C+Gilsinia+Lopez%2C+Alexandra+Olteanu%2C+Robert+Sim%2C+and+Hanna+Wallach%2E+2021%2E+Stereotyping+norwegian+salmon%3A+An+inventory+of+pitfalls+in+fairness+benchmark+datasets%2E+In+Proceedings+of+the+5)
5. Tolga Bolukbasi, Kai-Wei Chang, James Y Zou, Venkatesh Saligrama, and Adam T Kalai. 2016. [Man is to computer programmer as woman is to homemaker? debiasing word embeddings](https://md.research.elamur.ai/free/en/papers/bolukbasi-2016-man-computer-programmer-woman-homemaker/). Advances in neural information processing systems, 29.
6. Yuchen Cai, Ding Cao, Rongxi Guo, Yaqin Wen, Guiquan Liu, and Enhong Chen. 2024. Locating and mitigating gender bias in large language models. In International Conference on Intelligent Computing, pages 471–482. Springer. [find](https://scholar.google.com/scholar?q=Yuchen+Cai%2C+Ding+Cao%2C+Rongxi+Guo%2C+Yaqin+Wen%2C+Guiquan+Liu%2C+and+Enhong+Chen%2E+2024%2E+Locating+and+mitigating+gender+bias+in+large+language+models%2E+In+International+Conference+on+Intelligent+Computing%2C+pag)
7. Bhavik Chandna, Zubair Bashir, and Procheta Sen. 2025. Dissecting bias in llms: A mechanistic interpretability perspective. arXiv preprint arXiv:2506.05166. — [arXiv:2506.05166](https://arxiv.org/abs/2506.05166) [find](https://scholar.google.com/scholar?q=Bhavik+Chandna%2C+Zubair+Bashir%2C+and+Procheta+Sen%2E+2025%2E+Dissecting+bias+in+llms%3A+A+mechanistic+interpretability+perspective%2E+arXiv+preprint+arXiv%3A2506%2E05166%2E+%E2%80%94+%5BarXiv%3A2506%2E05166%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F)
8. Yuen Chen, Vethavikashini Chithrra Raghuram, Justus Mattern, Rada Mihalcea, and Zhijing Jin. 2024a. Causally testing gender bias in LLMs: A case study on occupational bias. In Causality and Large Models @NeurIPS 2024. [find](https://scholar.google.com/scholar?q=Yuen+Chen%2C+Vethavikashini+Chithrra+Raghuram%2C+Justus+Mattern%2C+Rada+Mihalcea%2C+and+Zhijing+Jin%2E+2024a%2E+Causally+testing+gender+bias+in+LLMs%3A+A+case+study+on+occupational+bias%2E+In+Causality+and+Large+Mode)
9. Yuheng Chen, Pengfei Cao, Yubo Chen, Kang Liu, and Jun Zhao. 2024b. Journey to the center of the knowledge neurons: Discoveries of language-independent knowledge neurons and degenerate knowledge neurons. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 17817–17825. [find](https://scholar.google.com/scholar?q=Yuheng+Chen%2C+Pengfei+Cao%2C+Yubo+Chen%2C+Kang+Liu%2C+and+Jun+Zhao%2E+2024b%2E+Journey+to+the+center+of+the+knowledge+neurons%3A+Discoveries+of+language%2Dindependent+knowledge+neurons+and+degenerate+knowledge+neuro)
10. Abhijith Chintam, Rahel Beloch, Willem Zuidema, Michael Hanna, and Oskar Van Der Wal. 2023. [Identifying and adapting transformer-components responsible for gender bias in an english language model](https://md.research.elamur.ai/free/en/papers/chintam-2023-identifying-adapting-transformer-components-responsible/). In Proceedings of the 6th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP, pa…
11. Arthur Conmy, Augustine Mavor-Parker, Aengus Lynch, Stefan Heimersheim, and Adrià Garriga-Alonso. 2023. Towards automated circuit discovery for mechanistic interpretability. Advances in Neural Information Processing Systems, 36:16318–16352. [find](https://scholar.google.com/scholar?q=Arthur+Conmy%2C+Augustine+Mavor%2DParker%2C+Aengus+Lynch%2C+Stefan+Heimersheim%2C+and+Adri%C3%A0+Garriga%2DAlonso%2E+2023%2E+Towards+automated+circuit+discovery+for+mechanistic+interpretability%2E+Advances+in+Neural+Informa)
12. Damai Dai, Li Dong, Yaru Hao, Zhifang Sui, Baobao Chang, and Furu Wei. 2021. Knowledge neurons in pretrained transformers. arXiv preprint arXiv:2104.08696. [find](https://scholar.google.com/scholar?q=Damai+Dai%2C+Li+Dong%2C+Yaru+Hao%2C+Zhifang+Sui%2C+Baobao+Chang%2C+and+Furu+Wei%2E+2021%2E+Knowledge+neurons+in+pretrained+transformers%2E+arXiv+preprint+arXiv%3A2104%2E08696%2E)
13. Xufeng Duan, Xinyu Zhou, Bei Xiao, and Zhenguang G Cai. 2024. [Unveiling language competence neurons: A psycholinguistic approach to model interpretability](https://md.research.elamur.ai/free/en/papers/duan-2025-unveiling-language-competence-neurons-psycholinguistic/). arXiv preprint arXiv:2409.15827.
14. Jacob Dunefsky and Arman Cohan. 2024. Observable propagation: Uncovering feature vectors in transformers. In Forty-first International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=Jacob+Dunefsky+and+Arman+Cohan%2E+2024%2E+Observable+propagation%3A+Uncovering+feature+vectors+in+transformers%2E+In+Forty%2Dfirst+International+Conference+on+Machine+Learning%2E)
15. Javier Ferrando and Elena Voita. 2024. [Information flow routes: Automatically interpreting language models at scale](https://md.research.elamur.ai/free/en/papers/ferrando-2024-information-flow-routes-automatically-interpreting/). arXiv preprint arXiv:2403.00824.
16. Danielle Gaucher, Justin Friesen, and Aaron C Kay. 2011. Evidence that gendered wording in job advertisements exists and sustains gender inequality. Journal of personality and social psychology, 101(1):109. [find](https://scholar.google.com/scholar?q=Danielle+Gaucher%2C+Justin+Friesen%2C+and+Aaron+C+Kay%2E+2011%2E+Evidence+that+gendered+wording+in+job+advertisements+exists+and+sustains+gender+inequality%2E+Journal+of+personality+and+social+psychology%2C+101%281)
17. Mor Geva, Roei Schuster, Jonathan Berant, and Omer Levy. 2021. Transformer feed-forward layers are key-value memories. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 5484–5495. [find](https://scholar.google.com/scholar?q=Mor+Geva%2C+Roei+Schuster%2C+Jonathan+Berant%2C+and+Omer+Levy%2E+2021%2E+Transformer+feed%2Dforward+layers+are+key%2Dvalue+memories%2E+In+Proceedings+of+the+2021+Conference+on+Empirical+Methods+in+Natural+Language+Pr)
18. Dirk Groeneveld, Iz Beltagy, Evan Walsh, Akshita Bhagia, Rodney Kinney, Oyvind Tafjord, Ananya Jha, Hamish Ivison, Ian Magnusson, Yizhong Wang, et al. 2024. Olmo: Accelerating the science of language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (… [find](https://scholar.google.com/scholar?q=Dirk+Groeneveld%2C+Iz+Beltagy%2C+Evan+Walsh%2C+Akshita+Bhagia%2C+Rodney+Kinney%2C+Oyvind+Tafjord%2C+Ananya+Jha%2C+Hamish+Ivison%2C+Ian+Magnusson%2C+Yizhong+Wang%2C+et+al%2E+2024%2E+Olmo%3A+Accelerating+the+science+of+language+)
19. Michael Hanna, Yonatan Belinkov, and Sandro Pezzelle. 2025. [Are formal and functional linguistic mechanisms dissociated in language models?](https://md.research.elamur.ai/free/en/papers/hanna-2026-are-formal-functional-linguistic-mechanisms/) Computational Linguistics, pages 1–40.
20. Michael Hanna, Sandro Pezzelle, and Yonatan Belinkov. 2024. Have faith in faithfulness: Going beyond circuit overlap when finding model mechanisms. In ICML 2024 Workshop on Mechanistic Interpretability. [find](https://scholar.google.com/scholar?q=Michael+Hanna%2C+Sandro+Pezzelle%2C+and+Yonatan+Belinkov%2E+2024%2E+Have+faith+in+faithfulness%3A+Going+beyond+circuit+overlap+when+finding+model+mechanisms%2E+In+ICML+2024+Workshop+on+Mechanistic+Interpretabilit)
21. Evan Hernandez, Arnab Sen Sharma, Tal Haklay, Kevin Meng, Martin Wattenberg, Jacob Andreas, Yonatan Belinkov, and David Bau. 2023. Linearity of relation decoding in transformer language models. arXiv preprint arXiv:2308.09124. [find](https://scholar.google.com/scholar?q=Evan+Hernandez%2C+Arnab+Sen+Sharma%2C+Tal+Haklay%2C+Kevin+Meng%2C+Martin+Wattenberg%2C+Jacob+Andreas%2C+Yonatan+Belinkov%2C+and+David+Bau%2E+2023%2E+Linearity+of+relation+decoding+in+transformer+language+models%2E+arXiv+)
22. Masahiro Kaneko and Danushka Bollegala. 2019. Gender-preserving debiasing for pre-trained word embeddings. arXiv preprint arXiv:1906.00742. [find](https://scholar.google.com/scholar?q=Masahiro+Kaneko+and+Danushka+Bollegala%2E+2019%2E+Gender%2Dpreserving+debiasing+for+pre%2Dtrained+word+embeddings%2E+arXiv+preprint+arXiv%3A1906%2E00742%2E)
23. Vivek Hruday Kavuri, Gargi Shroff, and Rahul Mishra. 2025. Neat: Concept driven neuron attribution in llms. arXiv preprint arXiv:2508.15875. — [arXiv:2508.15875](https://arxiv.org/abs/2508.15875) [find](https://scholar.google.com/scholar?q=Vivek+Hruday+Kavuri%2C+Gargi+Shroff%2C+and+Rahul+Mishra%2E+2025%2E+Neat%3A+Concept+driven+neuron+attribution+in+llms%2E+arXiv+preprint+arXiv%3A2508%2E15875%2E+%E2%80%94+%5BarXiv%3A2508%2E15875%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2508%2E15875%29)
24. Jong-Bok Kim and Ji-Young Ann. 2008. English tag questions: Corpus findings and theoretical implications. English Language and Linguistics, 25:103– 126. [find](https://scholar.google.com/scholar?q=Jong%2DBok+Kim+and+Ji%2DYoung+Ann%2E+2008%2E+English+tag+questions%3A+Corpus+findings+and+theoretical+implications%2E+English+Language+and+Linguistics%2C+25%3A103%E2%80%93+126%2E)
25. Maximilian Li and Lucas Janson. 2024. Optimal ablation for interpretability. Advances in Neural Information Processing Systems, 37:109233–109282. [find](https://scholar.google.com/scholar?q=Maximilian+Li+and+Lucas+Janson%2E+2024%2E+Optimal+ablation+for+interpretability%2E+Advances+in+Neural+Information+Processing+Systems%2C+37%3A109233%E2%80%93109282%2E)
26. Tomasz Limisiewicz and David Mareček. 2022. Don’t forget about pronouns: Removing gender bias in language models without losing factual gender information. arXiv preprint arXiv:2206.10744. [find](https://scholar.google.com/scholar?q=Tomasz+Limisiewicz+and+David+Marec%CC%8Cek%2E+2022%2E+Don%E2%80%99t+forget+about+pronouns%3A+Removing+gender+bias+in+language+models+without+losing+factual+gender+information%2E+arXiv+preprint+arXiv%3A2206%2E10744%2E)
27. Tomasz Limisiewicz, David Mareček, and Tomáš Musil. [Debiasing algorithm through model adaptation](https://md.research.elamur.ai/free/en/papers/limisiewicz-2023-debiasing-algorithm-through-model-adaptation/). In The Twelfth International Conference on Learning Representations. Tomasz Limisiewicz, David Mareček, and Tomáš Musil. 2025. Dual debiasing: Remove stereotypes and keep factual gender for fair langu…
28. Yan Liu, Yu Liu, Xiaokang Chen, Pin-Yu Chen, Daoguang Zan, Min-Yen Kan, and Tsung-Yi Ho. 2024. The devil is in the neurons: Interpreting and mitigating social biases in language models. In The Twelfth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Yan+Liu%2C+Yu+Liu%2C+Xiaokang+Chen%2C+Pin%2DYu+Chen%2C+Daoguang+Zan%2C+Min%2DYen+Kan%2C+and+Tsung%2DYi+Ho%2E+2024%2E+The+devil+is+in+the+neurons%3A+Interpreting+and+mitigating+social+biases+in+language+models%2E+In+The+Twelfth)
29. Yihong Liu, Runsheng Chen, Lea Hirlimann, Ahmad Dawar Hakimi, Mingyang Wang, Amir Hossein Kargaran, Sascha Rothe, François Yvon, and Hinrich Schütze. 2025. On relation-specific neurons in large language models. arXiv preprint arXiv:2502.17355. [find](https://scholar.google.com/scholar?q=Yihong+Liu%2C+Runsheng+Chen%2C+Lea+Hirlimann%2C+Ahmad+Dawar+Hakimi%2C+Mingyang+Wang%2C+Amir+Hossein+Kargaran%2C+Sascha+Rothe%2C+Fran%C3%A7ois+Yvon%2C+and+Hinrich+Sch%C3%BCtze%2E+2025%2E+On+relation%2Dspecific+neurons+in+large+langua)
30. Chris Mathwin, Guillaume Corlouer, Esben Kran, Fazl Barez, and Neel Nanda. 2023. Identifying a preliminary circuit for predicting gendered pronouns in gpt-2 small. URL: [https://itch.io/jam/mechint/rate/1889871](https://itch.io/jam/mechint/rate/1889871). [↗](https://itch.io/jam/mechint/rate/1889871)
31. Heiko Motschenbacher. 2016. A discursive approach to structural gender linguistics: theoretical and methodological considerations. Gender & Language, 10(2). [find](https://scholar.google.com/scholar?q=Heiko+Motschenbacher%2E+2016%2E+A+discursive+approach+to+structural+gender+linguistics%3A+theoretical+and+methodological+considerations%2E+Gender+%26+Language%2C+10%282%29%2E)
32. Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fiotto-Kaufman, Tal Haklay, Michael Hanna, et al. 2025. Mib: A mechanistic interpretability benchmark. arXiv preprint arXiv:2504.13151. — [arXiv:2504.13151](https://arxiv.org/abs/2504.13151) [find](https://scholar.google.com/scholar?q=Aaron+Mueller%2C+Atticus+Geiger%2C+Sarah+Wiegreffe%2C+Dana+Arad%2C+Iv%C3%A1n+Arcuschin%2C+Adam+Belfki%2C+Yik+Siu+Chan%2C+Jaden+Fiotto%2DKaufman%2C+Tal+Haklay%2C+Michael+Hanna%2C+et+al%2E+2025%2E+Mib%3A+A+mechanistic+interpretability+)
33. Yunzhi Yao, Ningyu Zhang, Zekun Xi, Mengru Wang, Ziwen Xu, Shumin Deng, and Huajun Chen. 2024. Knowledge circuits in pretrained transformers. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Yunzhi+Yao%2C+Ningyu+Zhang%2C+Zekun+Xi%2C+Mengru+Wang%2C+Ziwen+Xu%2C+Shumin+Deng%2C+and+Huajun+Chen%2E+2024%2E+Knowledge+circuits+in+pretrained+transformers%2E+In+The+Thirty%2Deighth+Annual+Conference+on+Neural+Informati)
34. Moin Nadeem, Anna Bethke, and Siva Reddy. 2020. [Stereoset: Measuring stereotypical bias in pretrained language models](https://md.research.elamur.ai/free/en/papers/nadeem-2021-stereoset-measuring-stereotypical-bias-pretrained/). arXiv preprint arXiv:2004.09456.
35. Zeping Yu and Sophia Ananiadou. 2024. Neuron-level knowledge attribution in large language models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 3267–3280. [find](https://scholar.google.com/scholar?q=Zeping+Yu+and+Sophia+Ananiadou%2E+2024%2E+Neuron%2Dlevel+knowledge+attribution+in+large+language+models%2E+In+Proceedings+of+the+2024+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+pages+3267)
36. Jingcheng Niu, Andrew Liu, Zining Zhu, and Gerald Penn. 2024. What does the knowledge neuron thesis have to do with knowledge? arXiv preprint arXiv:2405.02421. nostalgebraist. 2020. Interpreting gpt: The logit lens. LessWrong. — [arXiv:2405.02421](https://arxiv.org/abs/2405.02421) [find](https://scholar.google.com/scholar?q=Jingcheng+Niu%2C+Andrew+Liu%2C+Zining+Zhu%2C+and+Gerald+Penn%2E+2024%2E+What+does+the+knowledge+neuron+thesis+have+to+do+with+knowledge%3F+arXiv+preprint+arXiv%3A2405%2E02421%2E+nostalgebraist%2E+2020%2E+Interpreting+gpt%3A+)
37. Hadas Orgad and Yonatan Belinkov. 2022. Choose your lenses: Flaws in gender bias evaluation. In Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP), pages 151–167. [find](https://scholar.google.com/scholar?q=Hadas+Orgad+and+Yonatan+Belinkov%2E+2022%2E+Choose+your+lenses%3A+Flaws+in+gender+bias+evaluation%2E+In+Proceedings+of+the+4th+Workshop+on+Gender+Bias+in+Natural+Language+Processing+%28GeBNLP%29%2C+pages+151%E2%80%93167%2E)
38. Zeping Yu and Sophia Ananiadou. 2025. Understanding and mitigating gender bias in llms via interpretable neuron editing. arXiv preprint arXiv:2501.14457. — [arXiv:2501.14457](https://arxiv.org/abs/2501.14457) [find](https://scholar.google.com/scholar?q=Zeping+Yu+and+Sophia+Ananiadou%2E+2025%2E+Understanding+and+mitigating+gender+bias+in+llms+via+interpretable+neuron+editing%2E+arXiv+preprint+arXiv%3A2501%2E14457%2E+%E2%80%94+%5BarXiv%3A2501%2E14457%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F250)
39. Mahdi Zakizadeh, Kaveh Eskandari Miandoab, and Mohammad Taher Pilehvar. 2023. Difair: A benchmark for disentangled assessment of gender knowledge and bias. arXiv preprint arXiv:2310.14329. [find](https://scholar.google.com/scholar?q=Mahdi+Zakizadeh%2C+Kaveh+Eskandari+Miandoab%2C+and+Mohammad+Taher+Pilehvar%2E+2023%2E+Difair%3A+A+benchmark+for+disentangled+assessment+of+gender+knowledge+and+bias%2E+arXiv+preprint+arXiv%3A2310%2E14329%2E)
40. Jieyu Zhao, Yichao Zhou, Zeyu Li, Wei Wang, and Kai-Wei Chang. 2018. [Learning gender-neutral word embeddings](https://md.research.elamur.ai/free/en/papers/zhao-2018-learning-gender-neutral-word-embeddings/). arXiv preprint arXiv:1809.01496.
41. Hadas Orgad, Seraphina Goldfarb-Tarrant, and Yonatan Belinkov. 2022. How gender debiasing affects internal model representations, and why it matters. arXiv preprint arXiv:2204.06827. [find](https://scholar.google.com/scholar?q=Hadas+Orgad%2C+Seraphina+Goldfarb%2DTarrant%2C+and+Yonatan+Belinkov%2E+2022%2E+How+gender+debiasing+affects+internal+model+representations%2C+and+why+it+matters%2E+arXiv+preprint+arXiv%3A2204%2E06827%2E)
42. Christine A Smith, Ingrid Johnston-Robledo, Maureen C McHugh, and Joan C Chrisler. 2010. Words matter: The language of gender. Handbook of Gender Research in Psychology: Volume 1: Gender Research in General and Experimental Psychology, pages 361–377. [find](https://scholar.google.com/scholar?q=Christine+A+Smith%2C+Ingrid+Johnston%2DRobledo%2C+Maureen+C+McHugh%2C+and+Joan+C+Chrisler%2E+2010%2E+Words+matter%3A+The+language+of+gender%2E+Handbook+of+Gender+Research+in+Psychology%3A+Volume+1%3A+Gender+Research+in+G)
43. Mukund Sundararajan, Ankur Taly, and Qiqi Yan. 2017. Axiomatic attribution for deep networks. In International conference on machine learning, pages 3319– 3328. PMLR. [find](https://scholar.google.com/scholar?q=Mukund+Sundararajan%2C+Ankur+Taly%2C+and+Qiqi+Yan%2E+2017%2E+Axiomatic+attribution+for+deep+networks%2E+In+International+conference+on+machine+learning%2C+pages+3319%E2%80%93+3328%2E+PMLR%2E)
44. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. [Llama: Open and efficient foundation language models](https://md.research.elamur.ai/free/en/papers/touvron-2023-llama-open-efficient-foundation-language/). arXiv preprint arXiv:2302.13971.
45. A Vaswani. 2017. [Attention is all you need](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/). Advances in Neural Information Processing Systems.
46. Jesse Vig, Sebastian Gehrmann, Yonatan Belinkov, Sharon Qian, Daniel Nevo, Simas Sakenis, Jason Huang, Yaron Singer, and Stuart Shieber. 2020. Causal mediation analysis for interpreting neural nlp: The case of gender bias. arXiv preprint arXiv:2004.12265. — [arXiv:2004.12265](https://arxiv.org/abs/2004.12265) [find](https://scholar.google.com/scholar?q=Jesse+Vig%2C+Sebastian+Gehrmann%2C+Yonatan+Belinkov%2C+Sharon+Qian%2C+Daniel+Nevo%2C+Simas+Sakenis%2C+Jason+Huang%2C+Yaron+Singer%2C+and+Stuart+Shieber%2E+2020%2E+Causal+mediation+analysis+for+interpreting+neural+nlp%3A+Th)
47. Kevin Wang, Alexandre Variengien, Arthur Conmy, Buck Shlegeris, and Jacob Steinhardt. 2022. Interpretability in the wild: a circuit for indirect object identification in gpt-2 small. arXiv preprint arXiv:2211.00593. [find](https://scholar.google.com/scholar?q=Kevin+Wang%2C+Alexandre+Variengien%2C+Arthur+Conmy%2C+Buck+Shlegeris%2C+and+Jacob+Steinhardt%2E+2022%2E+Interpretability+in+the+wild%3A+a+circuit+for+indirect+object+identification+in+gpt%2D2+small%2E+arXiv+preprint+ar)

*Updated 1 November 2026*