# Indirect Prompt Injection Defense on Important Messages DeepSeek V3.2 (test) Source: [Indirect Prompt Injection Defense on Important Messages DeepSeek V3.2 (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:23:51.692Z Evaluation of model defense mechanisms against indirect prompt injection attacks using the test split of the Important Messages DeepSeek V3.2 dataset, measured by ASR, UA, and BU. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Indirect Prompt Injection Defense | | Dataset (including any stated split) | Important Messages DeepSeek V3.2 (test) | | Metric | ASR | | Unit | PERCENT | | Direction | Lower is better | | Results loaded | 1 of 1 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [ASR](): Lower is better; unit: PERCENT. - [UA](): Higher is better; unit: PERCENT. - [BU](): Higher is better; unit: PERCENT. ## Recorded results | Method | ASR | Result date | Report paper | | --- | --- | --- | --- | | No Defense | 42.68 | Not recorded | 6937acf86cfd488761a4e6a2 | ### No Defense ASR: 42.68. Result ID: 69962d5d5698665ec46edf67. [Reporting paper]() Recorded evaluation configuration: {"Model":"DeepSeek-V3.2"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.