计算机科学
布线(电子设计自动化)
人工智能
语义学(计算机科学)
钥匙(锁)
算法
数据挖掘
理论计算机科学
机器学习
算法设计
人工神经网络
数据建模
班级(哲学)
分布式计算
稳健性(进化)
作者
Chuancheng Shi,Shangze Li,Wenhua Wu,Hengzhi Chen,Fei Shen,Zhuo Li,Zhiyong Wang,Tat‐Seng Chua
标识
DOI:10.1109/tpami.2026.3747849
摘要
Despite their remarkable capabilities, large foundation models (LFMs) remain susceptible to adversarial manipulation. Current safety intervention methods primarily rely on the locality hypothesis, rigidly suppressing isolated neurons or features. However, harmful semantics manifest as distributed, cross-layer circuits, rendering localized ablations both brittle against complex jailbreaks and detrimental to general utility. To bridge this gap, we propose TraceRouter++, a routing-level safety intervention framework that explicitly traces safety-relevant causal routing pathways activated under harmful inputs and realigns their unsafe activation states. Our framework operates in a Discover-Trace-Realign paradigm: it first pinpoints a sensitive onset layer via attention divergence and disentangles dense activations into interpretable routing features using sparse autoencoders (SAEs). Subsequently, it maps these features to downstream causal circuits via feature influence scores (FIS). Crucially, departing from rigid suppression paradigms, TraceRouter++ introduces a constrained optimal transport (OT)-based soft routing realignment. By formulating safety intervention as a tractable coordinate- wise marginal optimal transport approximation, our method redirects unsafe activation marginals toward safe-reference marginals within causally identified routing subspaces while preserving non-selected coordinates at the intervention point. This mechanism safely redirects malicious semantic flows while leaving orthogonal computational routes intact. Extensive experiments across representative large language models (LLMs), multimodal large language models (MLLMs), and diffusion models (DMs) show that TraceRouter++ outperforms evaluated baselines, improving adversarial robustness while maintaining a favorable safety-utility trade-off.
科研通智能强力驱动
Strongly Powered by AbleSci AI