# What are the biases in my word embedding?

arXiv: https://arxiv.org/abs/1812.08769 · HTML: https://mlsanity.com/paper/1812.08769/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Dec 2018.

## Built on by (most cited first, 30 of 30)

- [Towards Understanding and Mitigating Social Biases in Language Models](https://mlsanity.com/paper/2106.13219/) (2021)
- [Gender bias and stereotypes in Large Language Models](https://mlsanity.com/paper/2308.14921/) (2023)
- [Almanac: Retrieval-Augmented Language Models for Clinical Medicine](https://mlsanity.com/paper/2303.01229/) (2023)
- [Do Language Models Know When They're Hallucinating References?](https://mlsanity.com/paper/2305.18248/) (2023)
- [Theories of "Gender" in NLP Bias Research](https://mlsanity.com/paper/2205.02526/) (2022)
- [SODAPOP: Open-Ended Discovery of Social Biases in Social Commonsense Reasoning Models](https://mlsanity.com/paper/2210.07269/) (2022)
- [Evaluating Biased Attitude Associations of Language Models in an Intersectional Context](https://mlsanity.com/paper/2307.03360/) (2023)
- [Hate Speech Classifiers Learn Human-Like Social Stereotypes](https://mlsanity.com/paper/2110.14839/) (2021)
- [WordBias: An Interactive Visual Tool for Discovering Intersectional Biases Encoded in Word Embeddings](https://mlsanity.com/paper/2103.03598/) (2021)
- ['Since Lawyers are Males..': Examining Implicit Gender Bias in Hindi Language Generation by LLMs](https://mlsanity.com/paper/2409.13484/) (2024)
- [Evaluating Metrics for Bias in Word Embeddings](https://mlsanity.com/paper/2111.07864/) (2021)
- [Social Norm Bias: Residual Harms of Fairness-Aware Algorithms](https://mlsanity.com/paper/2108.11056/) (2021)
- [Mind Your Bias: A Critical Review of Bias Detection Methods for Contextual Language Models](https://mlsanity.com/paper/2211.08461/) (2022)
- [More of the Same: Persistent Representational Harms Under Increased Representation](https://mlsanity.com/paper/2503.00333/) (2025)
- [Hire Me or Not? Examining Language Model's Behavior with Occupation Attributes](https://mlsanity.com/paper/2405.06687/) (2024)
- [Statistical Uncertainty in Word Embeddings: GloVe-V](https://mlsanity.com/paper/2406.12165/) (2024)
- [AI and Blackness: Towards moving beyond bias and representation](https://mlsanity.com/paper/2111.03687/) (2021)
- [Representation Bias of Adolescents in AI: A Bilingual, Bicultural Study](https://mlsanity.com/paper/2408.01961/) (2024)
- [Deconstructing Depression Stigma: Integrating AI-driven Data Collection and Analysis with Causal Knowledge Graphs](https://mlsanity.com/paper/2502.06075/) (2025)
- [An investigation of structures responsible for gender bias in BERT and DistilBERT](https://mlsanity.com/paper/2401.06495/) (2024)
- [Protected group bias and stereotypes in Large Language Models](https://mlsanity.com/paper/2403.14727/) (2024)
- [Counterfactual Augmentation for Multimodal Learning Under Presentation Bias](https://mlsanity.com/paper/2305.14083/) (2023)
- [SenteCon: Leveraging Lexicons to Learn Human-Interpretable Language Representations](https://mlsanity.com/paper/2305.14728/) (2023)
- [Character-level Chinese Backpack Language Models](https://mlsanity.com/paper/2310.12751/) (2023)
- [Implicit Priors Editing in Stable Diffusion via Targeted Token Adjustment](https://mlsanity.com/paper/2412.03400/) (2024)
- [Human-Imitating Metrics for Training and Evaluating Privacy Preserving Emotion Recognition Models Using Sociolinguistic Knowledge](https://mlsanity.com/paper/2104.08792/) (2021)
- [Investigating Bias In Automatic Toxic Comment Detection: An Empirical Study](https://mlsanity.com/paper/2108.06487/) (2021)
- [The SAME score: Improved cosine based bias score for word embeddings](https://mlsanity.com/paper/2203.14603/) (2022)
- [A Holistic Indicator of Polarization to Measure Online Sexism](https://mlsanity.com/paper/2404.02205/) (2024)
- [On Debiasing Text Embeddings Through Context Injection](https://mlsanity.com/paper/2410.12874/) (2024)

## Abstract (arXiv)

This paper presents an algorithm for enumerating biases in word embeddings. The algorithm exposes a large number of offensive associations related to sensitive features such as race and gender on publicly available embeddings, including a supposedly "debiased" embedding. These biases are concerning in light of the widespread use of word embeddings. The associations are identified by geometric patterns in word embeddings that run parallel between people's names and common lower-case tokens. The algorithm is highly unsupervised: it does not even require the sensitive features to be pre-specified. This is desirable because: (a) many forms of discrimination--such as racial discrimination--are linked to social constructs that may vary depending on the context, rather than to categories with fixed definitions; and (b) it makes it easier to identify biases against intersectional groups, which depend on combinations of sensitive features. The inputs to our algorithm are a list of target tokens, e.g. names, and a word embedding. It outputs a number of Word Embedding Association Tests (WEATs) that capture various biases present in the data. We illustrate the utility of our approach on publicly available word embeddings and lists of names, and evaluate its output using crowdsourcing. We also show how removing names may not remove potential proxy bias.