LongPIBench: A Long-Context Benchmark for Prompt Injection
尚未评估计划受阻研究发现claim-detection-tradeoff

The paper reports that detection-based defenses exhibit an extreme false-positive/false-negative trade-off on long-context inputs, with some methods flagging benign inputs and others missing attacks.

来源:paper:pp. 9-10, Sections 5.1 and 5.4, Table 3; Appendix F

报告指标与观测值

false_positive_rate

m-datasentinel-fpr-paper-review

论文报告 0.43 fraction

实际观测 — fraction

Assessment(0)

这条 Claim 暂无已发布的不可变 Assessment。

关联运行(0)

这条 Claim 暂无关联执行。