# SysBench: Can Large Language Models Follow System Messages?

arXiv: https://arxiv.org/abs/2408.10943 · HTML: https://mlsanity.com/paper/2408.10943/

## Key facts

- Cited by 28 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 37 references that are in the map.
- First posted on arXiv 20 Aug 2024.

## Builds on (most cited first, 28 of 28)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [WebGPT: Browser-assisted question-answering with human feedback](https://mlsanity.com/paper/2112.09332/) (2021)
- [AgentBench: Evaluating LLMs as Agents](https://mlsanity.com/paper/2308.03688/) (2023)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model](https://mlsanity.com/paper/2405.04434/) (2024)
- [ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools](https://mlsanity.com/paper/2406.12793/) (2024)
- [Large Language Models Can Be Easily Distracted by Irrelevant Context](https://mlsanity.com/paper/2302.00093/) (2023)
- [The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions](https://mlsanity.com/paper/2404.13208/) (2024)
- [Orca: Progressive Learning from Complex Explanation Traces of GPT-4](https://mlsanity.com/paper/2306.02707/) (2023)
- [ERNIE 3.0: Large-scale Knowledge Enhanced Pre-training for Language Understanding and Generation](https://mlsanity.com/paper/2107.02137/) (2021)
- [TALM: Tool Augmented Language Models](https://mlsanity.com/paper/2205.12255/) (2022)
- [In-Context Impersonation Reveals Large Language Models' Strengths and Biases](https://mlsanity.com/paper/2305.14930/) (2023)
- [Benchmarking Complex Instruction-Following with Multiple Constraints Composition](https://mlsanity.com/paper/2407.03978/) (2024)
- [Investigating Cultural Alignment of Large Language Models](https://mlsanity.com/paper/2402.13231/) (2024)
- [InFoBench: Evaluating Instruction Following Ability in Large Language Models](https://mlsanity.com/paper/2401.03601/) (2024)
- [FollowBench: A Multi-level Fine-grained Constraints Following Benchmark for Large Language Models](https://mlsanity.com/paper/2310.20410/) (2023)
- [Aligning to Thousands of Preferences via System Message Generalization](https://mlsanity.com/paper/2405.17977/) (2024)
- [FOFO: A Benchmark to Evaluate LLMs' Format-Following Capability](https://mlsanity.com/paper/2402.18667/) (2024)
- [Can Large Language Models Understand Real-World Complex Instructions?](https://mlsanity.com/paper/2309.09150/) (2023)
- [CFBench: A Comprehensive Constraints-Following Benchmark for LLMs](https://mlsanity.com/paper/2408.01122/) (2024)
- [Can LLMs Follow Simple Rules?](https://mlsanity.com/paper/2311.04235/) (2023)
- [Instruction-following Evaluation through Verbalizer Manipulation](https://mlsanity.com/paper/2307.10558/) (2023)
- [Is the System Message Really Important to Jailbreaks in Large Language Models?](https://mlsanity.com/paper/2402.14857/) (2024)
- [Beyond ChatBots: ExploreLLM for Structured Thoughts and Personalized Model Responses](https://mlsanity.com/paper/2312.00763/) (2023)
- [SoFA: Shielded On-the-fly Alignment via Priority Rule Following](https://mlsanity.com/paper/2402.17358/) (2024)

## Built on by (most cited first, 28 of 28)

- [BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games](https://mlsanity.com/paper/2411.13543/) (2024)
- [AdaCoT: Pareto-Optimal Adaptive Chain-of-Thought Triggering via Reinforcement Learning](https://mlsanity.com/paper/2505.11896/) (2025)
- [IHEval: Evaluating Language Models on Following the Instruction Hierarchy](https://mlsanity.com/paper/2502.08745/) (2025)
- [AGENTIF: Benchmarking Instruction Following of Large Language Models in Agentic Scenarios](https://mlsanity.com/paper/2505.16944/) (2025)
- [LIFBench: Evaluating the Instruction Following Performance and Stability of Large Language Models in Long-Context Scenarios](https://mlsanity.com/paper/2411.07037/) (2024)
- [SOPBench: Evaluating Language Agents at Following Standard Operating Procedures and Constraints](https://mlsanity.com/paper/2503.08669/) (2025)
- [Position is Power: System Prompts as a Mechanism of Bias in Large Language Models (LLMs)](https://mlsanity.com/paper/2505.21091/) (2025)
- [Prompt Injection as Role Confusion](https://mlsanity.com/paper/2603.12277/) (2026)
- [WebGames: Challenging General-Purpose Web-Browsing AI Agents](https://mlsanity.com/paper/2502.18356/) (2025)
- [Security Considerations for Artificial Intelligence Agents](https://mlsanity.com/paper/2603.12230/) (2026)
- [IF-VidCap: Can Video Caption Models Follow Instructions?](https://mlsanity.com/paper/2510.18726/) (2025)
- [Baichuan Alignment Technical Report](https://mlsanity.com/paper/2410.14940/) (2024)
- [Reverse Preference Optimization for Complex Instruction Following](https://mlsanity.com/paper/2505.22172/) (2025)
- [System Message Generation for User Preferences using Open-Source Models](https://mlsanity.com/paper/2502.11330/) (2025)
- [LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues](https://mlsanity.com/paper/2507.13681/) (2025)
- [Analyzing and Internalizing Complex Policy Documents for LLM Agents](https://mlsanity.com/paper/2510.11588/) (2025)
- [OmniCap-IF: Benchmarking and Improving Instruction Following Abilities for Omni-Video Captioning](https://mlsanity.com/paper/2606.08572/) (2026)
- [Steering Instruction Hierarchies at Inference Time](https://mlsanity.com/paper/2607.26228/) (2026)
- [SafePyramid: A Hierarchical Benchmark for In-context Policy Guardrailing](https://mlsanity.com/paper/2606.29887/) (2026)
- [Efficient Safety Alignment of Language Models via Latent Personality Traits](https://mlsanity.com/paper/2607.07918/) (2026)
- [How LLMs Follow Instructions: Skillful Coordination, Not a Universal Mechanism](https://mlsanity.com/paper/2604.06015/) (2026)
- [System-Prompt Anchoring with Cross-Attention Layers](https://mlsanity.com/paper/2605.09737/) (2026)
- [@skills: Attention is all you have](https://mlsanity.com/paper/2608.12610/) (2026)
- [Compliance, Capability, and Conflict: Benchmarking Multimodal LLMs under System Messages](https://mlsanity.com/paper/2608.19207/) (2026)
- [Seeing Through Conflicts: Improving Instruction Hierarchy Alignment in Vision-Language Models](https://mlsanity.com/paper/2609.22234/) (2026)
- [RLVR2: Reinforcement Learning with Verifiable Rubric-based Ranking](https://mlsanity.com/paper/2609.23457/) (2026)
- [Hunyuan-A13B Technical Report](https://mlsanity.com/paper/2609.27284/) (2026)
- [AgentPersonaBench: Benchmarking Persona-Driven User Simulation](https://mlsanity.com/paper/2610.04379/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have become instrumental across various applications, with the customization of these models to specific scenarios becoming increasingly critical. System message, a fundamental component of LLMs, is consist of carefully crafted instructions that guide the behavior of model to meet intended goals. Despite the recognized potential of system messages to optimize AI-driven solutions, there is a notable absence of a comprehensive benchmark for evaluating how well LLMs follow system messages. To fill this gap, we introduce SysBench, a benchmark that systematically analyzes system message following ability in terms of three limitations of existing LLMs: constraint violation, instruction misjudgement and multi-turn instability. Specifically, we manually construct evaluation dataset based on six prevalent types of constraints, including 500 tailor-designed system messages and multi-turn user conversations covering various interaction relationships. Additionally, we develop a comprehensive evaluation protocol to measure model performance. Finally, we conduct extensive evaluation across various existing LLMs, measuring their ability to follow specified constraints given in system messages. The results highlight both the strengths and weaknesses of existing models, offering key insights and directions for future research. The open source library SysBench is available at https://github.com/PKU-Baichuan-MLSystemLab/SysBench.