mlsanityData to 14 Oct 2026

LLMs Silently Correct African American English: Auditing and Mitigating Dialect Bias via Activation Steering

arXiv 2607.06845 ↗First posted 7 Jul 2026cs.CLcited by 0

On the map: Mechanistic interpretability → Probing deception, bias and beliefs

Abstract (arXiv)

African American English (AAE), a rule-governed dialect spoken by over 30 million people, is routinely misinterpreted and "corrected" by large language models (LLMs). Across six instruction-tuned LLMs (14B to 70B), we show that state-of-the-art models systematically prefer Standard American English (SAE) continuations even when the preceding context is in AAE, effectively rewriting AAE into SAE. We present an end-to-end framework to audit and mitigate this bias. For auditing, we introduce conditional Dialect Group Invariance (cDGI), which isolates true model bias from translator-induced artifacts, and a feature-level localization analysis that identifies which AAE markers most strongly trigger bias; we find that syntactic constructions, especially negative concord (e.g., "ain't nobody"), are universal triggers across all models. For mitigation, we introduce, to our knowledge, the first application of activation steering to dialect bias: a training-free, test-time method that extracts dialect directions via causal tracing and injects them into bias-relevant layers. Activation steering reduces bias 5 to 20 times more than prompting while preserving SAE fluency. To enable this work, we release REAL-AAE , the largest real-AAE parallel corpus to date: 17,479 AAE/SAE/ AAE_back triplets from natural tweets (2 to 6 times larger than prior real-AAE resources), validated automatically (BERTScore F1 = 0.95) and by three native AAE speakers (83.0% semantic agreement).

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Probing deception, bias and beliefs, one of 13 directions in Mechanistic interpretability. Builds on 19 of its 31 references that are in the map. First posted on arXiv 7 Jul 2026.

Builds on

19 of 31 references.

References · 19

  1. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  2. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  3. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,917
  4. 2023Steering Llama 2 via Contrastive Activation Additioncited by 1,085
  5. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
  6. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 345
  7. 2021Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detectioncited by 264
  8. 2023In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steeringcited by 245
  9. 2019Racial Bias in Hate Speech and Abusive Language Detection Datasetscited by 209
  10. 2024Linguistic Bias in ChatGPT: Language Models Reinforce Dialect Discriminationcited by 56
  11. 2022Multi-VALUE: A Framework for Cross-Dialectal English NLPcited by 53
  12. 2022VALUE: Understanding Dialect Disparity in NLUcited by 47
7 more
  1. 2023Evaluation of African American Language Bias in Natural Language Generationcited by 36
  2. 2024Assessing Dialect Fairness and Robustness of Large Language Models in Reasoning Taskscited by 22
  3. 2025Rejected Dialects: Biases Against African American Language in Reward Modelscited by 20
  4. 2024AAVENUE: Detecting LLM Biases on NLU Tasks in AAVE via a Novel Benchmarkcited by 11
  5. 2024Predicting the Target Word of Game-playing Conversations using a Low-Rank Dialect Adapter for Decoder Modelscited by 6
  6. 2025EnDive: A Cross-Dialect Benchmark for Fairness and Performance in Large Language Modelscited by 5
  7. 2025DialectGen: Benchmarking and Improving Dialect Robustness in Multimodal Generationcited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Disparities in LLM Reasoning Accuracy and Explanations: A Case Study on African American English
  2. 2025Activation Steering for Bias Mitigation: An Interpretable Approach to Safer LLMs
  3. 2026Evaluating the Usage of African-American Vernacular English in Large Language Models
  4. 2025Bias Testing and Mitigation in Black Box LLMs using Metamorphic Relations
  5. 2025Can Small-Scale Data Poisoning Exacerbate Dialect-Linked Biases in Large Language Models?
  6. 2026The Dialect Tax: Dialectal Biases Persist throughout the Language Modeling Pipeline
  7. 2026Dialect vs Demographics: Quantifying LLM Bias from Implicit Linguistic Signals vs. Explicit User Profiles
  8. 2026The Wrong Kind of Right: Quantifying and Localizing Misfired Alignment in LLMs