PPL
Fan Yang
Papers (4)
STAR-PólyaMath: Multi-Agent Reasoning under Persistent Meta-Strategic SupervisionWhen Reasoning Traces Become Performative: Step-Level Evidence that Chain-of-Thought Is an Imperfect Oversight ChannelA Single Layer to Explain Them All:Understanding Massive Activations in Large Language ModelsTrapping Attacker in Dilemma: Examining Internal Correlations and External Influences of Trigger for Defending GNN Backdoors