Showing 2026Show all
2 papers · 1 filter
cs.LG2026
To Cache or Not to Cache: Understanding and Mitigating Safety Incompatibility in Cross-Model KV Cache Reuse
Zhiwei Lin, Jiahao Liu, Yan Liu +1
Cross-model KV cache sharing reduces prefill computation by allowing a receiver model to reuse prefix KV states generated by a related sender model. Existing methods optimize such…
cs.AI2026
Inverting the Shield: Systematically Generating Safety Tests from Policy Specifications
Xiaoyue Lu, Xianglin Yang, Haijun Liu +4
The widespread integration of Large Language Models (LLMs) necessitates rigorous and systematic safety evaluation. Existing paradigms either rely on constructed benchmarks to asses…