Search papers, labs, and topics across Lattice.
2
0
6
0
Proprietary MLLMs generate visually appealing code from Figma designs, but fall short on crucial aspects like responsive layouts and maintainability, revealing a surprising gap in their practical utility.
DeepSeek LLMs are surprisingly vulnerable to prompt injection attacks that combine semantic and character-level mutations, outperforming single-strategy attacks by 12.5% in misuse success rate.
