Back to AI News
arxiv

SPADE Uses Soft Prompts as a Signal for AI-Text Detection

Researchers introduce SPADE, a framework that detects AI-generated text by measuring how well a prompt explains a candidate text using a Pointwise Mutual Information term. SPADE trains one soft prompt per class and predicts the class whose prompt best explains the text. The method is designed to improve robustness and transfer across datasets, with experiments reporting state-of-the-art out-of-distribution performance against various detectors.

AI Brief

It offers a new detection approach that uses prompt influence rather than only text features, which may improve robustness in open-world AI-text detection.

arXiv:2611.16929v1 Announce Type: cross Abstract: The easy access to Large Language Models (LLMs) has made AI-generated texts on the web increasingly ubiquitous, making AI-text detection a crucial task. Prompts greatly influence the generation process, but existing detectors focus solely on the generated texts --- supervised methods learn from labeled texts only, and zero-shot methods rely on the text's probability distributions --- ignoring the underlying generation process, which limits their robustness in open-world scenarios. We introduce SPADE: Soft Prompts for Ai-text DEtection, a novel framework for detecting AI-generated text that transfers better across datasets. Specifically, we measure the prompt's influence and use it as a discriminative signal. We characterize the prompt's influence by introducing a Pointwise Mutual Information (PMI) term that captures how much a prompt explains a text. Based on this quantity, SPADE trains one soft prompt per class to amplify the separation, and predicts the class whose prompt best explains the candidate text. Extensive experiments on several benchmarks against various detectors demonstrate the robustness and effectiveness of our method, achieving state-of-the-art (SOTA) performance on out-of-distribution (OOD) datasets.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!