Metraipolis › archive › dc_d6521e7eefdb
Acmite: Mitigating Gender Bias in LLMs through Concept-Guided Mutual Information — arXiv:2610.01696
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.01696
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T13:55:57.132Z
- SHA-256 of the archived text
72f03a5eb167e2671f6e9e7c8d7a412cf35f27650efe4f76e6395bf2879664ed- Stable URL of this page
https://metraipolis.com/documento/dc_d6521e7eefdb
How to cite
Acmite: Mitigating Gender Bias in LLMs through Concept-Guided Mutual Information — arXiv:2610.01696 — arXiv (en), https://arxiv.org/abs/2610.01696 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d6521e7eefdb (Metraipolis archive dc_d6521e7eefdb, sha256 72f03a5eb167e2671f6e9e7c8d7a412cf35f27650efe4f76e6395bf2879664ed)
BibTeX
@misc{dcd6521e7eefdb,
title = {Acmite: Mitigating Gender Bias in LLMs through Concept-Guided Mutual Information — arXiv:2610.01696},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.01696},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d6521e7eefdb; sha256 72f03a5eb167e2671f6e9e7c8d7a412cf35f27650efe4f76e6395bf2879664ed},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.01696. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) can reproduce social stereotypes from their training data, motivating extensive research on model debiasing. However, existing methods often rely on explicit biased examples or predefined group-term substitutions, making them sensitive to wording and less effective at capturing stereotype concepts shared across diverse contexts. More importantly, they typically suppress biased outputs without explicitly modeling the statistical dependence between model outputs and the underlying stereotype concepts. We propose Acmite, a lightweight concept-guided framework for targeted and selective debiasing. Acmite represents stereotypes as structured semantic concepts and uses maximal marginal relevance (MMR) to select diverse concepts for debiasing. Inspired by mutual information minimization, it approximates this dependence with token-level KL divergence while preserving task semantics. A lightweight LoRA adapter is trained with the base model frozen and activated at inference time only when the input is sufficiently similar to stereotype-related concepts; otherwise, the original model is used directly. We evaluate Acmite on BBQ, CrowS-Pairs, and StereoSet, and assess general capability preservation on ARC-Challenge, GSM8K, and PIQA. Experiments across three LLMs show that Acmite effectively mitigates gender bias across complementary evaluation formats while maintaining competitive performance on bias-unrelated tasks. Anonymous code and data are available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.01696 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-06T12:15:34.741Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).