$k$NNProxy: Efficient Training-Free Proxy Alignment for Black-Box Zero-Shot LLM-Generated Text Detection

$k$NNProxy: Efficient Training-Free Proxy Alignment for Black-Box Zero-Shot LLM-Generated Text Detection explores A training-free framework for efficient and robust detection of LLM-generated text by aligning black-box proxy models.. Commercial viability score: 7/10 in LLM Security.