arXiv:2603.13683v1 Announce Type: cross
Abstract: Although debiased LLMs perform well on known bias patterns, they often fail to generalize to unfamiliar bias prompts, producing toxic outputs. We first validate that such high-bias prompts constitute a emphdistribution shift via OOD detection, and show static models degrade under this shift. To adapt on-the-fly, we propose textbfCAP-TTA, a test-time adaptation framework that performs context-aware LoRA updates only when the bias-risk emphtrigger exceeds a threshold, using a precomputed diagonal emphpreconditioner for fast and stable updates. Across toxic-prompt settings and benchmarks, CAP-TTA reduces bias (confirmed by human evaluation) while achieving much lower update latency than AdamW/SGD; it also mitigates catastrophic forgetting by significantly improving narrative fluency over SOTA debiasing baseline while maintaining comparable debiasing effectiveness.
Translating AI research into reality: summary of the 2025 voice AI Symposium and Hackathon
The 2025 Voice AI Symposium represented a transition from conceptual research to clinical implementation in vocal biomarker science. Hosted by the NIH-funded Bridge2AI-Voice consortium, the