What it measures
Evaluates multi-turn adherence to complex developer system prompts under distracting or conflicting user messages.
Reading the measurement
Separate satisfying instructions from factual accuracy, human preference and safe behavior. Inspect strict all-constraint success as well as average criterion credit. Judge choice and adversarial prompt selection can materially change the result.
Published evaluation
Instruction satisfaction rate under system-message constraints; CSR and multi-turn SSR are distinct source metrics.
Where this fits in an evaluation suite
Format compliance, preference, truthfulness, and safety require separate measurements. Keep desirable writing style from masking factual errors.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Instruction satisfaction rate under system-message constraints; CSR and multi-turn SSR are distinct source metrics.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
- SysBench: Can Large Language Models Follow System Messages? — published result table ↗
All 12 authors
Yanzhao Qin, Tao Zhang, Tao Zhang, Yanjun Shen, Wenjing Luo, Haoze Sun, Yan Zhang, Yujing Qiao, Weipeng Chen, Zenan Zhou, Wentao Zhang, Bin Cui
Reviewed 2026-10-26. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.