Search papers, labs, and topics across Lattice.
École Polytechnique, IP Paris, CNRS), AMIAD (Agence Ministérielle pour l'IA de Défense)
3
0
6
LLM safety patches can appear effective on benchmarks but cause severe collateral damage to benign behaviors, a critical flaw missed by aggregate metrics.
Prompt injection defenses can degrade LLM agent utility by altering benign behavior; RAISED achieves robustness without this cost through self-distillation on generated scenarios.
Forget prompt engineering – surgically altering a model's internal activations can jailbreak it, exposing vulnerabilities even when the input looks harmless.
