Paper introduces a new algorithm to detect LLM-generated text.
problem Detecting LLM-generated text to prevent misinformation.
method Adaptively learns the distance between original and rewritten text.
result Empirically, the new algorithm outperforms existing methods in most scenarios.