Research map / Large language models / Language, safety and interpretability
Assessing Social and Intersectional Biases in Contextualized Word Representations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Measure bias with contextual-word encodings rather than relying on sentence encodings alone | Contextual word models · Embedding association tests | significant-effect instances detected: 34 only with c-word; 24 only with sent; 35 with both | Mixed |
| Compare racial associations with gender associations | Contextual word models · Embedding association tests | proportion of significant positive effect sizes | Mixed |
| Compare the combined African American and female identity against the most privileged group | Intersectional tests · Intersectional identities | effect size | Helped |
| Compare race against gender as the constituent minority identity | Intersectional tests · Intersectional identities | number of significant positive associations: 9 vs 2 | Helped |
| Increase model size | BERT and GPT families · Overall association tests | proportion of significant positive effect sizes: blc: 0.48, bbc: 0.33; GPT: 0.3, GPT-2_117M: 0.32, GPT-2_345M: 0.23 | Mixed |
Closest papers by meaning
- Are Models Biased on Text without Gender-related Language? (ICLR 2024 · similarity 0.89)
- SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examples (CVPR 2024 · similarity 0.88)
- SocialStigmaQA: A Benchmark to Uncover Stigma Amplification in Generative Language Models (AAAI 2024 · similarity 0.88)
- More of the Same: Persistent Representational Harms Under Increased Representation (NeurIPS 2025 · similarity 0.87)
- Biases Mitigation and Expressiveness Preservation in Language Models: A Comprehensive Pipeline (Student Abstract) (AAAI 2024 · similarity 0.86)
- FairImagen: Post-Processing for Bias Mitigation in Text-to-Image Models (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.