Many-Tier Instruction Hierarchy in LLM Agents

TL;DR

This paper introduces ManyIH, a new paradigm and benchmark for resolving instruction conflicts across many privilege levels in large language model agents, highlighting current models' limitations.

Contribution

The paper proposes ManyIH, a scalable instruction hierarchy framework, and introduces ManyIH-Bench, a benchmark with realistic, multi-level conflicting instructions for LLM agents.

Findings

Current models perform poorly (~40% accuracy) on ManyIH-Bench.

ManyIH-Bench includes 853 tasks with up to 12 conflicting instruction levels.

The benchmark spans 46 real-world agent scenarios.

Abstract

Large language model agents receive instructions from many sources-system messages, user prompts, tool outputs, other agents, and more-each carrying different levels of trust and authority. When these instructions conflict, agents must reliably follow the highest-privilege instruction to remain safe and effective. The dominant paradigm, instruction hierarchy (IH), assumes a fixed, small set of privilege levels (typically fewer than five) defined by rigid role labels (e.g., system > user). This is inadequate for real-world agentic settings, where conflicts can arise across far more sources and contexts. In this work, we propose Many-Tier Instruction Hierarchy (ManyIH), a paradigm for resolving instruction conflicts among instructions with arbitrarily many privilege levels. We introduce ManyIH-Bench, the first benchmark for ManyIH. ManyIH-Bench requires models to navigate up to 12 levels of conflicting instructions with varying privileges, comprising 853 agentic tasks (427 coding and 426 instruction-following). ManyIH-Bench composes constraints developed by LLMs and verified by humans to create realistic and difficult test cases spanning 46 real-world agents. Our experiments show that even the current frontier models perform poorly (~40% accuracy) when instruction conflict scales. This work underscores the urgent need for methods that explicitly target fine-grained, scalable instruction conflict resolution in agentic settings.