IHDec: Divergence-Steered Contrastive Decoding for Securing Multi-Turn Instruction Hierarchies

arXiv:2606.29960 · cs.CL · Submitted 2026-06-29 · Read on arXiv

cs.CL

Submitted: 2026-06-29

Updated: 2026-09-17

Comments: EMNLP 2026 Findings

Code: https://github.com/nxcolelxu/IHDec

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Instruction Hierarchy — 21× in this paper
• Accuracy — 9× in this paper · explained in 1 episode
• Alignment — 9× in this paper · explained in 1 episode
• Benchmark — 8× in this paper · explained in 2 episodes
• Jensen-Shannon Divergence — 6× in this paper
• Core — 5× in this paper · explained in 2 episodes
• Steering — 5× in this paper · explained in 2 episodes
• Ablation — 4× in this paper · explained in 1 episode
• Embeddings — 4× in this paper · explained in 2 episodes
• Fine-Tuning — 3× in this paper · explained in 5 episodes
• Large Language Models (LLMs) — 3× in this paper · explained in 6 episodes
• Robustness — 3× in this paper · explained in 5 episodes
• Ablation study — 2× in this paper · explained in 1 episode
• Alignment Tax — 2× in this paper · explained in 4 episodes
• Consistency — 2× in this paper · explained in 3 episodes
• Conversational Memory — 2× in this paper
• LLM-as-a-Judge — 2× in this paper · explained in 3 episodes
• Adaptability — 1× in this paper · explained in 1 episode
• Adversarial Context — 1× in this paper · explained in 1 episode
• Agentic System — 1× in this paper · explained in 2 episodes
• Alignment Methods — 1× in this paper · explained in 1 episode
• GRPO — 1× in this paper · explained in 6 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Hallucination — 1× in this paper · explained in 4 episodes
• Hallucination Mitigation — 1× in this paper · explained in 2 episodes
• Hierarchical Framework — 1× in this paper · explained in 2 episodes
• Inference-Time Steering — 1× in this paper · explained in 2 episodes
• Interpretability — 1× in this paper · explained in 15 episodes
• KV cache — 1× in this paper · explained in 3 episodes
• Model-Agnostic — 1× in this paper · explained in 2 episodes
• Multi-Agent System — 1× in this paper · explained in 6 episodes
• Multi-agent Collaboration — 1× in this paper · explained in 2 episodes
• Specific Instruction — 1× in this paper
• Structural Constraints — 1× in this paper · explained in 2 episodes

Sources

• Toward General Instruction-Following Alignment for Retrieval-Augmented Generation
• Steer Model beyond Assistant: Controlling System Prompt Strength via Contrastive Decoding
• IH-Challenge: A Training Dataset to Improve Instruction Hierarchy on Frontier LLMs
• Stronger Enforcement of Instruction Hierarchy via Augmented Intermediate Representations
• Lost in the Middle: How Language Models Use Long Contexts
• The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
• Diagnose, Localize, Align: A Full-Stack Framework for Reliable LLM Multi-Agent Systems under Instruction Conflicts
• Hierarchical Alignment: Enforcing Hierarchical Instruction-Following in LLMs through Logical Consistency
• Reasoning Up the Instruction Ladder for Controllable Language Models

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering