Search papers, labs, and topics across Lattice.
8
0
9
14
AMPLE-Math, a reusable suite of 5,319 mathematical problems with six reasoning views that share the same answer, is constructed and AMPLE-Math, a reusable suite of 5,319 mathematical problems with six reasoning views that share the same answer, is compared, suggesting that OPSD can improve access to existing reasoning capabilities through parameters shared by direct-response and thinking-enabled inference.
Safety compliance can vary by over 28 points among LLMs with similar predictive accuracy, revealing hidden risks in flight predictions.
LMSM reduces attack success rates by over 90% while maintaining nearly full throughput, revolutionizing how we secure LLM deployments.
RePolicy achieves superior safety-policy invocation in language model agents, adapting dynamically to changing contexts and unseen trajectories.
Cognitive traps in LLM memory can lead to over 10% performance degradation, challenging the assumption that more memory always improves reasoning.
Mechanist uncovers a surprising safety risk where unsafe traits can transfer across modalities, challenging assumptions about training data safety.
SafeCA slashes jailbreak success rates by 20% while adding virtually no latency, revolutionizing defenses for text-to-video models.
The shift from parameter-centric to system-level adaptation in continual learning could redefine how we build and interact with AI models.
