Arc Sentry Crushes LLM Guard in Prompt Injection Detection: 92% vs 70%

I built Arc Sentry, a pre-generation prompt injection detector for open-weight LLMs. Instead of scanning text for patterns after the fact, it reads the model's internal residual stream before generate() is called and blocks requests that destabilize the model's information geometry. Head to head benchmark on a 130-prompt SaaS deployment dataset: Arc Sentry: 92% detection, 0% false positives LLM Guard: 70% detection, 3.3% false positives The difference is architectural. LLM Guard classifies inputs after they enter the model, but Arc Sentry intercepts threats at the neural level before generation even starts.