PACT achieves perfect security and utility under oracle provenance by enforcing argument-level trust contracts based on semantic roles and cross-step provenance tracking, outperforming invocation-level monitors in AgentDojo evaluations.
Instructional Segment Embedding: Improving LLM Safety with Instruction Hierarchy
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 7verdicts
UNVERDICTED 7roles
background 1polarities
background 1representative citing papers
ManyIH and ManyIH-Bench address instruction conflicts in LLM agents with up to 12 privilege levels across 853 tasks, revealing frontier models achieve only ~40% accuracy.
Prompt injection defenses create a security-fidelity tradeoff with no model or defense achieving both high security and high fidelity on the SecFid benchmark across 1,168 examples.
Empirical test on real news histories shows unconstrained LLM reranking amplifies conspiratorial content exposure while prompt constraints can increase ideological diversity with limited relevance cost.
Black-box optimization outperforms gradient-based methods for prompt injection on LLM agents, with success depending on attacker model strength and limited transfer from small to frontier models.
The paper argues that agent security is best addressed as a systems problem by applying principles from operating systems, networks, and formal methods rather than relying solely on model robustness improvements.
Frontier AI agents introduce new confidentiality, integrity, and availability risks through changed assumptions on code-data separation and authority boundaries, requiring layered defenses like sandboxing and policy enforcement.
citing papers explorer
-
LLM-Assisted Reranking to Operationalize Nuanced Objectives in Recommender Systems
Empirical test on real news histories shows unconstrained LLM reranking amplifies conspiratorial content exposure while prompt constraints can increase ideological diversity with limited relevance cost.