3 papers
cs.CR2026
Stronger Judges Do Not See More: Deployment-Integrity Blindness in LLM Evaluation
Jianchang Su, Dawei Xiang, Yifan Zhang +1
LLM-agent evaluation reports behavioral scores but leaves the identity of the deployed runtime stack unreported. A tampered deployment that preserves evaluated outputs passes every…
cs.LG2026
SAIR: Cost-Efficient Multi-Stage ML Pipeline Autoscaling via In-Context Reinforcement Learning
Jianchang Su, Yifan Zhang, Shengkai Lin +4
Multi-stage ML inference pipelines are difficult to autoscale due to heterogeneous resources, cross-stage coupling, and dynamic bottleneck migration. We present SAIR, an autoscalin…
cs.DC2025
ExpertFlow: Adaptive Expert Scheduling and Memory Coordination for Efficient MoE Inference
Zixu Shen, Kexin Chu, Yifan Zhang +3
The expansion of large language models is increasingly limited by the constrained memory capacity of modern GPUs. To mitigate this, Mixture-of-Experts (MoE) architectures activate…