AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Uncovering and Quantifying Social Biases in Code Generation

Yan Liu, Xiaokang Chen, Yan Gao, Zhe Su, Fengji Zhang, Daoguang Zan, Jian-Guang Lou, Pin-Yu Chen, Tsung-Yi Ho

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: Bias measurement and mitigation · 4 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
scale InCoder from 1.3B to 6.7BInCoder · automatic evaluationCBS Tot. (%): 23.52 -> 32.55Hurt
scale CodeGen-Mono from 350M to 2.7BCodeGen-Mono · automatic evaluationCBS Tot. (%): 9.36 -> 45.15Hurt
scale CodeGen-Mono from 2.7B to 6.1BCodeGen-Mono · automatic evaluationCBS Tot. (%): 45.15 -> 62.65Hurt
scale InCoder from 1.3B to 6.7BInCoder · HumanEvalPass@1 (%): 9.00 -> 15.20Improved
scale CodeGen-Mono from 350M to 6.1BCodeGen-Mono · HumanEvalPass@1 (%): 12.76 -> 26.13Improved
scale CodeGen-Mono from 350M to 6.1BCodeGen-Mono · human evaluationCBS Tot. (%): 5.69 -> 68.25Hurt
increase number of prompt functions from 1 to 3InCoder · prompt studyCBS (%): 8.53 -> 40.01Hurt
increase number of prompt functions from 1 to 3CodeGen · prompt studyCBS (%): 14.05 -> 52.63Hurt
make prompt functions human-relevant (0Y2N to 2Y0N)InCoder · prompt studyCBS (%): 31.55 -> 55.75Hurt
make prompt functions human-relevant (0Y2N to 2Y0N)CodeGen · prompt studyCBS (%): 39.30 -> 63.42Hurt
place human-relevant function near signature (NEG to POS order)InCoder · prompt studyCBS (%): 48.38 -> 49.05Hurt
place human-relevant function near signature (NEG to POS order)CodeGen · prompt studyCBS (%): 56.91 -> 58.26Hurt

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