Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
2
0
3
2
Models already know when a question is fundamentally unanswerable—they hallucinate anyway because their internal "impossibility" representation is completely orthogonal to their safety-refusal circuitry.
VLMs, typically praised for their multimodal synergy, can be easily weaponized to manipulate search rankings via imperceptible image perturbations and fluent textual suffixes, outperforming unimodal attacks.