PPL
Jiacheng Liu
Papers (5)
MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual ContextsLLMSurgeon: Diagnosing Data Mixture of Large Language ModelsIndexMem: Learned KV-Cache Eviction with Latent Memory for Long-Context LLM InferenceFocused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video DiffusionVisMMOE: Exploiting Visual-Expert Affinity for Efficient Visual-Language MoE Offloading