paper-with-me

Papers

IHEval: Evaluating Language Models on Following the Instruction Hierarchy

2025-02-12 · Zhihan Zhang, Shiyang Li, Zixuan Zhang, Xin Liu, Haoming Jiang, Xianfeng Tang, Yifan Gao, Zheng Li, Haodong Wang, Zhaoxuan Tan, Yichuan Li, Qingyu Yin, Bing Yin, Meng Jiang

The instruction hierarchy, which establishes a priority order from system messages to user messages, conversation history, and tool outputs, is essential for ensuring consistent and safe behavior in language models (LMs). Despite its importance, this topic receives limited attention, and there is a lack of comprehensive benchmarks for evaluating models' ability to follow the instruction hierarchy. We bridge this gap by introducing IHEval, a novel benchmark comprising 3,538 examples across nine tasks, covering cases where instructions in different priorities either align or conflict. Our evaluation of popular LMs highlights their struggle to recognize instruction priorities. All evaluated models experience a sharp performance decline when facing conflicting instructions, compared to their original instruction-following performance. Moreover, the most competitive open-source model only achieves 48% accuracy in resolving such conflicts. Our results underscore the need for targeted optimization in the future development of LMs.

📄 PDF Abstract BibTeX arXiv:2502.08745

Code (1)

ytyz1307zzh/IHEval 공식 구현

Tasks

Instruction Following

Methods 이 논문이 사용한 방법론

ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

Where Instruction Hierarchy Breaks: Diagnosing and Repairing Failures in Reasoning Language Models

2026-06-05 · Sanjay Kariyappa, G. Edward Suh arxiv

Reasoning language models deployed in agentic workflows must follow an instruction hierarchy: when instructions from different sources conflict, the model should obey the highest-privilege applicable instruction. Existin…

M-IFEval: Multilingual Instruction-Following Evaluation

2025-02-07 · Antoine Dussolle, Andrea Cardeña Díaz, Shota Sato, Peter Devine

Instruction following is a core capability of modern Large language models (LLMs), making evaluating this capability essential to understanding these models. The Instruction Following Evaluation (IFEval) benchmark from t…

Instruction Following

Reasoning Up the Instruction Ladder for Controllable Language Models

2025-10-30 · Zishuo Zheng, Vidhisha Balachandran, Chan Young Park, Faeze Brahman 외 arxiv

As large language model (LLM) based systems take on high-stakes roles in real-world decision-making, they must reconcile competing instructions from multiple sources within a single prompt context. Enforcing an instructi…

Reinforcement LearningInstruction Following

Instructional Segment Embedding: Improving LLM Safety with Instruction Hierarchy

2024-10-09 · Tong Wu, Shujian Zhang, Kaiqiang Song, Silei Xu 외

Large Language Models (LLMs) are susceptible to security and safety threats, such as prompt injection, prompt extraction, and harmful requests. One major cause of these vulnerabilities is the lack of an instruction hiera…

Instruction Following

Empowering Reliable Visual-Centric Instruction Following in MLLMs

2026-01-06 · Weilei He, Feng Ju, Zhiyuan Fan, Rui Min 외 arxiv

Evaluating the instruction-following (IF) capabilities of Multimodal Large Language Models (MLLMs) is essential for rigorously assessing how faithfully model outputs adhere to user-specified intentions. Nevertheless, exi…

Instruction Following