AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Investigating Gender Bias in Language Models Using Causal Mediation Analysis

Jesse Vig, Sebastian Gehrmann, Yonatan Belinkov, Sharon Qian, Daniel Nevo, Yaron Singer, Stuart Shieber

NeurIPS 2020 · Architectures and efficiency › Attention and transformers · approach: Language model pretraining and induction · 133 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (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 changedWhereMeasuredEffect
pretraining vs random initializationGPT2-small · Winobiastotal effect of gender bias: 0.066 -> 0.249Improved
increase model size (small to xl)GPT2 · Winobiastotal effect of gender bias: 0.249 -> 1.049Improved
increase model sizeGPT2 · Professionstotal effect of gender biasMixed
intervene on only top 10 attention heads instead of all 144GPT2-small mediation analysis · Winobiasindirect effectNo clear effect
intervene on top 4% of neurons instead of allGPT2-small mediation analysis · Professionsindirect effectNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy