Bo Gao

The paper proposes SALO, a novel detector that monitors the dynamic, layer-wise activation pattern (Refusal Trajectory) to improve jailbreak detection robustness compared to traditional methods relying on static terminal representations.

The paper proposes PCDM, a diffusion-based framework that enables highly stealthy and effective data poisoning attacks against Federated Learning systems, significantly degrading global performance while evading detection.

The paper introduces CV-Arena, a large-scale open benchmark for instructional computer vision, demonstrating that professional-grade image editing requires advanced capabilities in physical reasoning and structural control.

The paper introduces Hierarchical Adaptive Budgeter (HAB), a framework that improves LLM reasoning efficiency by adaptively allocating computational resources to match the intrinsic complexity of both problems and individual reasoning steps.