Search papers, labs, and topics across Lattice.
3
0
8
0
Multimodal models are surprisingly vulnerable to subtle text-based backdoor attacks using common words, making them far more practical and stealthy than previous visual trigger methods.
A multi-agent system can dramatically improve the accuracy of LLM-based housing consultations, jumping from 75% to 95% correctness in real-world scenarios.
Achieve controllable and scalable speech generation with MOSS-TTS, enabling zero-shot voice cloning and long-form synthesis.
