DeceptPrompt: Exploiting LLM-driven Code Generation via Adversarial Natural Language Instructions

Last updated