Search papers, labs, and topics across Lattice.
3
0
7
3
Even advanced LLMs fail 83% of the time on a novel 3D spatial reasoning benchmark, revealing a critical gap between 2D perception and precise 3D action control.
Merging aligned models can inadvertently expose them to jailbreak attacks rooted in their shared pretrained foundations, with a new method that exploits this vulnerability showing high transfer success rates.
A single ranking can adapt to any frame budget, improving accuracy and reducing latency without retraining the model.