PragLocker protects agent prompts as IP by building non-portable obfuscated versions that function only on the intended LLM through code-symbol semantic anchoring followed by target-model feedback noise injection.
arXiv preprint arXiv:2312.00025 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CR 2years
2026 2representative citing papers
An attack aligns differently shuffled intermediate activations from secure Transformer inference queries to recover model weights with low error using roughly one dollar of queries.
citing papers explorer
-
PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts
PragLocker protects agent prompts as IP by building non-portable obfuscated versions that function only on the intended LLM through code-symbol semantic anchoring followed by target-model feedback noise injection.
-
On the (In-)Security of the Shuffling Defense in the Transformer Secure Inference
An attack aligns differently shuffled intermediate activations from secure Transformer inference queries to recover model weights with low error using roughly one dollar of queries.