Search papers, labs, and topics across Lattice.
5
6
10
Control interventions are often detected by LLMs, with awareness levels varying significantly across models and tasks, revealing vulnerabilities in AI safety protocols.
LLMs struggle with structured 2D tasks when inputs are serialized into 1D, revealing a surprising performance gap compared to vision-augmented models that directly process the 2D layout.
DNA can now encode entirely new-to-nature enzymatic reactions, thanks to a generative model that designs proteins around user-specified chemistries.
Mimicking human cognition, FLAIR lets dialogue models "think while listening," boosting performance without adding latency.
A global consensus on AI safety risks and capabilities has emerged from a panel of 100+ independent experts, representing a landmark effort in international collaboration.
