This study proposes a Query-Guided Cross-Attention CNN (QGC-CNN) framework that combines the local feature extraction power of CNNs with transformer-based global attention using learnable query tokens ...

Model checkpoint variant: Our experiments use the Qwen-7B-Chat variant with grouped-query attention (GQA), which has a smaller effective KV-cache footprint than a full-size 7B dense model at the ...

Five Queries Are Enough: Query-Efficient and Surrogate-Free Membership Inference Attacks on RAG via Entailment Nguyen Linh Bao Nguyen, Wanlun Ma, Viet Vo, Alsharif Abuadbba, Minghong Fang, Jun Zhang, ...

The project automatically fetches the latest papers from arXiv based on keywords. The subheadings in the README file represent the search keywords. Only the most recent articles for each keyword are ...