Linear probes on pre-generation hidden states detect indirect prompt-injection exposure across six agentic LLMs with 90%+ AUROC, and probe-gated anti-injection reasoning reduces attack success toward zero.
Ignore previous prompt: Attack techniques for language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Your Agentic LLMs Secretly Encode Latent Signals of Indirect Prompt-Injection Exposure
Linear probes on pre-generation hidden states detect indirect prompt-injection exposure across six agentic LLMs with 90%+ AUROC, and probe-gated anti-injection reasoning reduces attack success toward zero.