MMODELYST
People/Hao Guo
PPL

Hao Guo

Papers (3)
When Mean CE Fails: Median CE Can Better Track Language Model QualityBeyond Inference-Only Deployment: Comparing Weight-Based Consolidation Against Cascading CompactionCompiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost