12 papers
SlotGuard: Stop Oversharing Private Local Context in LLM Agent Transcri
Haocheng Xia, Yongjoo Park
LLM agents can leak privacy (e.g., paths, emails) and credentials (e.g., API keys) as agent observations (e.g., tool outputs, shell logs, and file reads) are appended to provider-b…
LazyAttention: Efficient Retrieval-Augmented Generation with Deferred Positional Encoding
Haocheng Xia, Mihir Pamnani, Hanxi Fang +2
Key-value (KV) caching accelerates inference of large language models (LLMs) by reusing past computations for generated tokens. Its importance becomes even greater in long-context…
MojoFrame: Dataframe Library in Mojo Language
Shengya Huang, Zhaoheng Li, Derek Warner +1
Mojo is an emerging programming language built on MLIR (Multi-Level Intermediate Representation) and supports JIT (Just-in-Time) compilation. It enables transparent hardware-specif…
Cloud-Native Vector Search: A Comprehensive Performance Analysis
Zhaoheng Li, Wei Ding, Silu Huang +5
Vector search has been widely employed in recommender system and retrieval-augmented-generation pipelines, commonly performed with vector indexes to efficiently find similar items…
Chipmink: Efficient Delta Identification for Massive Object Graph
Supawit Chockchowwat, Sumay Thakurdesai, Zhaoheng Li +2
Ranging from batch scripts to computational notebooks, modern data science tools rely on massive and evolving object graphs that represent structured data, models, plots, and more.…
QStore: Quantization-Aware Compressed Model Storage
Raunak Shah, Zhaoheng Li, Yongjoo Park
Modern applications commonly leverage large, multi-modal foundation models. These applications often feature complex workflows that demand the storage and usage of similar models i…