Safety
GGSS: Geodesic-Gated Spherical Steering for Inference-Time Debiasing of Generative Vision-Language Models
arXiv:2608.25375v1 Announce Type: cross Abstract: Generative vision-language models (VLMs) are increasingly used in human-centered settings, yet they can produce demographically biased outputs even wh
arXiv:2608.25375v1 Announce Type: cross Abstract: Generative vision-language models (VLMs) are increasingly used in human-centered settings, yet they can produce demographically biased outputs even when images differ only in controlled attributes such as perceived race or gender. However, existing inference-time debiasers were largely designed for static embeddings or CLIP-like models rather than generative VLMs. We propose GGSS---Geodesic-Gated Spherical Steering---a norm-preserving intervention that discovers a counterfactual bias subspace on the unit hypersphere, steers visual tokens along geodesic arcs, and uses an adaptive gate to focus correction on tokens that carry stronger demographic signal. We evaluate four generative VLMs against ten adapted inference-time debiasing baselines and prompt-based mitigation under a single operating-point protocol across categorical, pairwise, and occupation-gender bias tests, while also measuring general visual-language capability. GGSS achieves the lowest average bias on all four models, significant on three of four backbones under paired permutation tests, while preserving MMStar accuracy within +/- 0.6 p.p. of the unsteered baseline. Code is available at https://github.com/dukesun99/GGSS.
Related
- Universal Boosts, Specific Suppressors: Sparse Autoencoder Steering of Medical Vision-Language Models
- Selective Test-Time Debiasing for CLIP via Reward Gating
- Decomposed On-Policy Distillation for Vision-Language Reasoning: Steering Gradients for Visual Grounding
Source: arXiv cs.CL | 2026-08-27