ExaGPT: Example-Based Machine-Generated Text Detection for Human Interpretability

TL;DR

ExaGPT is an interpretable LLM-generated text detection method that uses span similarity to human-written and machine-generated texts, significantly improving detection accuracy and human decision support.

Contribution

It introduces a human-inspired span similarity approach for detecting LLM-generated texts, enhancing interpretability and accuracy over existing methods.

Findings

ExaGPT outperforms prior detectors by up to +37.0 points in accuracy.

Providing span examples improves human judgment of detection correctness.

The method is effective across four domains and three generator types.

Abstract

Detecting texts generated by Large Language Models (LLMs) could cause grave mistakes due to incorrect decisions, such as undermining students' academic dignity. LLM text detection thus needs to ensure the interpretability of the decision, which can help users judge how reliably correct its prediction is. When humans verify whether a text is human-written or LLM-generated, they intuitively investigate which of them it shares more similar spans with. However, existing interpretable detectors are not aligned with the human decision-making process and fail to offer evidence that users easily understand. To bridge this gap, we introduce ExaGPT, an interpretable detection approach grounded in the human decision-making process for verifying the origin of a text. ExaGPT identifies a text by checking whether it shares more similar spans with human-written vs. with LLM-generated texts from a datastore. This approach can provide similar span examples that contribute to the decision for each span in the text as evidence. Our human evaluation demonstrates that providing similar span examples contributes more effectively to judging the correctness of the decision than existing interpretable methods. Moreover, extensive experiments in four domains and three generators show that ExaGPT massively outperforms prior interpretable detectors by up to +37.0 points of accuracy at a false positive rate of 1%.