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

Prevention-based defenses that perform well on short-context benchmarks retain substantial attack success on the paper's long-context synthetic benchmark, although PromptLocate and MetaSecAlign 8B are lower than several simpler defenses on some tasks.

来源:paper:pp. 10-11, Sections 5.3-5.4, Tables 4-5

报告指标与观测值

attack_success_rate

m-metsecalign-paper-review

论文报告 1 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。