The paper examines how large language models (LLMs) respond to different demographic cues—such as names—when users seek advice, focusing on race and gender in a U.S. context. It finds that using different cues for the same group leads to only partially overlapping changes in model responses, producing inconsistent conclusions about personalization and unstable bias metrics. The authors argue that LLMs react to linguistic signals tied to cues rather than to stable demographic categories, and they call for evaluations that use multiple cues and consider underlying mechanisms.
By Manuel Tonneau, Neil K. R. Sehgal, Niyati Malhotra, Sharif Kazemi, Victor Orozco-Olvera, Ana Mar\'ia Mu\~noz Boudet, Lakshmi Subramanian, Samuel P. Fraiberger, Sharath Chandra Guntuku, Valentin Hofmann
arXiv:2604. 01925v2 Announce Type: replace-cross Abstract: Large Language Models increasingly suppress biased outputs when demographic identity is stated explicitly, yet may still exhibit implicit biases when identity is conveyed indirectly.
By Bhaskara Hanuma Vedula, Darshan Anghan, Ishita Goyal, Ponnurangam Kumaraguru, Abhijnan Chakraborty
arXiv:2605.31556v2 Announce Type: replace-cross
Abstract: Alignment teaches vision-language models (VLMs) to avoid expressing demographic biases, and when gender is clearly visible they largely succe...
By Arnau Marin-Llobet, Simon Henniger, Mahzarin R. Banaji
arXiv:2501. 02211v2 Announce Type: replace-cross Abstract: Large language models (LLMs) reproduce homogeneity bias -- the tendency to portray marginalized groups as more internally similar than dominant groups -- but whether this bias is stable or an artifact of inference settings has only been studied in single proprietary models.
By Messi H. J. Lee
arXiv:2501. 02211v3 Announce Type: replace-cross Abstract: Large language models (LLMs) reproduce homogeneity bias -- the tendency to portray marginalized groups as more internally similar than dominant groups -- but whether this bias generalizes across models, is stable under different inference settings, or depends on how group identity is signaled remains unstudied.
By Messi H. J. Lee
arXiv:2606. 14117v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly evaluated for bias using adaptations of human psychological paradigms, yet methodological limitations-particularly the conflation of refusal behavior with task performance-have hindered clear interpretation.
By Achraf Cohen, Andrew Kincaid
arXiv:2508. 03483v3 Announce Type: replace-cross Abstract: While prior research on text-to-image generation has predominantly focused on biases in human depictions, demographic bias in generated objects remains relatively underexplored.
By Dasol Choi, Jihwan Lee, Minjae Lee, Minsuk Kahng
The paper introduces a unified framework that simultaneously measures intrinsic (encoded) and extrinsic (expressed) gender bias in large language models using identical neutral prompts. It finds a consistent link between latent gender information and output bias, but shows that alignment via supervised fine‑tuning reduces expressed bias while leaving internal gender associations largely intact and reactivatable by adversarial prompts. The study also demonstrates that debiasing gains on structured benchmarks may not transfer to realistic tasks such as story generation.
By Nour Bouchouchi, Thibault Laugel, Xavier Renard, Christophe Marsala, Marie-Jeanne Lesot, Marcin Detyniecki
arXiv:2608. 11735v1 Announce Type: cross Abstract: Large language models (LLMs) often shift their outputs in response to implicit demographic cues even when users never state a demographic identity.
By Yueru Yan, Siqi Wu, Thai Le
The paper examines a specific type of hallucination in large language models caused by spurious correlations—unintended, statistically prominent associations in training data such as surnames linked to nationalities. These hallucinations are confidently produced, persist regardless of model scaling or refusal fine‑tuning, and evade existing detection methods like confidence filtering and inner‑state probing. The authors use controlled synthetic experiments and evaluations on both open‑source and proprietary LLMs, including GPT‑5, to demonstrate the failure of current detection techniques and provide a theoretical explanation for why statistical biases undermine confidence‑based approaches.
By Shaowen Wang, Yiqi Dong, Ruinian Chang, Tansheng Zhu, Yuebo Sun, Kaifeng Lyu, Jian Li
arXiv:2610.23587v1 Announce Type: new
Abstract: Post-training is widely assumed to reduce the social biases language models inherit from pretraining, yet direct evidence for this assumption remains m...
By Manthan Gadhia, Alexander Hoyle, Karolina Sta\'nczak
arXiv:2601.12868v2 Announce Type: replace-cross
Abstract: Large language models (LLMs) increasingly operate in high-stakes settings where demographic attributes such as race and ethnicity may be expl...
By Shiyue Hu, Ruizhe Li, Yanjun Gao