CORTEXA
← Browse

Wang Kang

2 papers indexed

arxivcs.AR2026-07-31

Selective KV Cache Protection for Noise-Resilient LLM Inference on Analog Compute-In-Memory Systems

Yuannuo Feng, Wenyong Zhou, Yuang Ma, Yizhe Chen, Wenshuai Yao, Yuxin Xie, et al.

Analog compute-in-memory (CIM) arrays have emerged as a promising substrate for energy-efficient LLM inference, particularly for weight-stationary computations in linear layers. However, extending analog CIM to attention mechanisms introduces a fundamental challenge: KV cache ope…

View free PDFSource page
arxivcs.AR2026-07-26

CircuitWeave: Topology-Behavior Alignment for Executable Multimodal RTL Generation

Jiahao Feng, Haiyan Qin, Zhiwei Xie, Wang Kang

Text-only LLMs generate RTL from natural-language specifications, but prose can leave connectivity, register boundaries, and state-output relations implicit even when interfaces and cycle-level behavior are specified. Schematics can make these structural relations explicit and th…

View free PDFSource page