- Rating
- 1
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper proposes AlignSentinel to detect an input only when the instruction it carries overrides the higher-priority instruction. Compared with existing works, AlignSentinel distinguishes misaligned instructions, aligned instructions, and non-instruction inputs and flags only the first category. AlignSentinel includes an open-weight detection model on attention features between the two input regions. The paper conducts evaluations on an LLM-generated benchmark with 8 domains for direct and indirect injection. The results show that AlignSentinel achieves almost zero false-positive and false-negative rates and has powerful generalization on IHEval and OpenPromptInjection.
Critical strengths
-
The problem formulation addresses an evaluation gap between benign instructions and malicious injections. Testing whether a detector distinguishes them from conflicting instructions is informative and effective.
-
Comprehensive evaluations are conducted on self-curated benchmarks and existing benchmarks.
-
The proposed AlignSentinel is simple yet effective.
Critical weaknesses
-
I can hardly agree with the paper's arguments on "Existing detection defenses typically classify any input containing an instruction as malicious" and "existing prompt injection benchmarks do not account for instruction hierarchy." Nearly all prompt injection papers consider instruction hierarchy for their benchmark curation. I cannot see any difference between your curated benchmark and existing benchmarks.
-
The proposed benchmark lacks construction details and its synthetic construction may be hacked via training. The agent and domain splits share a similar generator and construction recipe. And the generation process deliberately creates aligned, opposite, and extended constraints. I cannot see anything new or novel in the construction.
-
Weak baselines are compared. The performance of frontier LLMs is missing.
-
Overall, the paper lacks novelty and comprehensive baseline comparisons. Attention-based detection itself is already established by AttnTracker and the benchmark has nothing new.
Questions for authors
Please refer to my listed weaknesses, especially Weakness 1.