Search papers, labs, and topics across Lattice.
3
0
7
2
This work introduces PACT (Pressure-Applied Compliance Testing), a benchmark for rule-following under pressure in AI agents assisting employees in daily tasks across twelve regulated enterprise domains and forty-eight scenarios, each set in a realistic multi-turn conversation.
Up to 50.2% of unauthorized requests can be falsely authorized by LLM memory, with executors blindly acting on these permissions 98.6% of the time.
Satellites can shed tasks and balance workloads without explicit policies or routing protocols, thanks to a novel decentralized runtime that uses a marginal cost signal derived from battery levels and other constraints.
