Six state-of-the-art LLMs systematically prefer Standard American English over AAE continuations, and a training-free activation steering method reduces this bias 5-20x more than prompting while preserving fluency.
Racial Bias in Hate Speech and Abusive Language Detection Datasets
3 Pith papers cite this work, alongside 43 external citations. Polarity classification is still indexing.
3
Pith papers citing it
43
external citations · OpenAlex
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 3roles
background 1polarities
background 1representative citing papers
BBQ is a new benchmark dataset showing that QA models often default to social stereotypes, achieving up to 3.4 points higher accuracy when the correct answer aligns with bias.
Automated hate speech detectors show poor alignment with heterogeneous in-group judgments on reclaimed slur usage, driven by low inter-annotator agreement and contextual features like derogatory intent.
citing papers explorer
-
LLMs Silently Correct African American English: Auditing and Mitigating Dialect Bias via Activation Steering
Six state-of-the-art LLMs systematically prefer Standard American English over AAE continuations, and a training-free activation steering method reduces this bias 5-20x more than prompting while preserving fluency.
-
BBQ: A Hand-Built Bias Benchmark for Question Answering
BBQ is a new benchmark dataset showing that QA models often default to social stereotypes, achieving up to 3.4 points higher accuracy when the correct answer aligns with bias.
-
IYKYK (But AI Doesn't): Automated Content Moderation Does Not Capture Communities' Heterogeneous Attitudes Towards Reclaimed Language
Automated hate speech detectors show poor alignment with heterogeneous in-group judgments on reclaimed slur usage, driven by low inter-annotator agreement and contextual features like derogatory intent.