PPL
Junhao Dong
Papers (5)
Adaptive Probe-based Steering for Robust LLM JailbreakingCATA: Continual Machine Unlearning via Conflict-Averse Task ArithmeticICED: Concept-level Machine Unlearning via Interpretable Concept DecompositionTSNN: A Non-parametric and Interpretable Framework for Traffic Time Series ForecastingBiAxisAudit: A Novel Framework to Evaluate LLM Bias Across Prompt Sensitivity and Response-Layer Divergence