← Home

Technical Reports

• [HPDIC26-08a] ATLAS: Accelerator-local Temporal Locality Analysis and Selection for Mixture-of-Experts LLM Inference. DOI: 10.5281/zenodo.21782950
• [HPDIC26-07c] STALK: Adaptive Candidate and Refinement Execution for GPU--DRAM Vector Search. DOI: 10.5281/zenodo.21714858
• [HPDIC26-07b] Post-Hoc Embedding Refinement with a Single Anchor Negative. DOI: 10.5281/zenodo.21540555
• [HPDIC26-07a] Label-Free Self-Distillation for Token Reduction in Frozen Multimodal Models. DOI: 10.5281/zenodo.21540301