PPL
Yilun Yao
Papers (4)
ConMoE: Expert-Pool Consolidation via Prototype Reassignment for MoE CompressionHarness-Bench: Measuring Harness Effects across Models in Realistic Agent WorkflowsMemAudit: Post-hoc Auditing of Poisoned Agent Memory via Causal Attribution and Structural Anomaly DetectionFormal Skill: Programmable Runtime Skills for Efficient and Accurate LLM Agents