{"slug": "dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text", "title": "DWT-Fusion: A Signal-Based Framework for Training-Free LLM-Generated Text Detection", "summary": "Researchers introduced DWT-Fusion, a training-free signal-based framework for detecting LLM-generated text using discrete wavelet analysis of token-level log-probability sequences. The best single wavelet configurations achieved AUROC values of 0.9872, 0.8185, and 0.7138 on HC3, M4, and MAGE datasets, respectively, while calibration-weighted voting improved AUROC to 0.9919, 0.8477, and 0.7471. The framework provides effective and interpretable detection signals without requiring supervised training.", "body_md": "arXiv:2607.22026v1 Announce Type: new\nAbstract: Detecting LLM-generated text remains challenging under zero-shot and training-free conditions, especially when detectors must generalize across datasets, domains, and unseen generators. While existing training-free approaches exploit language-model statistics as detection signals, they typically characterize a text through global measures that summarize overall model behavior. Consequently, potentially informative local and multiscale variations in token-level predictability may remain underutilized. Motivated by this observation, we introduce DWT-Fusion, a training-free signal-based framework for detecting LLM-generated text using discrete wavelet analysis of token-level log-probability sequences produced by a proxy causal language model. The proposed framework analyzes these sequences through wavelet-based multiresolution signal representations and derives detection signals from localized probability dynamics. We further evaluate four training-free voting variants, including equal-weight hard voting, equal-weight soft voting, calibration-weighted hard voting, and calibration-weighted soft voting, to combine multiple wavelet configurations without training a supervised meta-classifier. We evaluate the framework on HC3, M4, and MAGE using GPT-Neo-2.7B, GPT-J-6B, Falcon-7B, and LLaMA-3-8B as proxy models. The best single wavelet configurations achieve AUROC values of 0.9872, 0.8185, and 0.7138 on HC3, M4, and MAGE, respectively. With calibration-weighted voting, the best ensemble variants further improve AUROC to 0.9919, 0.8477, and 0.7471. These findings show that DWT-based multiresolution scoring and calibration-guided voting fusion provide effective and interpretable signals for training-free LLM-generated text detection.", "url": "https://wpnews.pro/news/dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text", "canonical_source": "https://arxiv.org/abs/2607.22026", "published_at": "2026-07-27 04:00:00+00:00", "updated_at": "2026-07-27 04:25:21.841777+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-research"], "entities": ["DWT-Fusion", "HC3", "M4", "MAGE", "GPT-Neo-2.7B", "GPT-J-6B", "Falcon-7B", "LLaMA-3-8B"], "alternates": {"html": "https://wpnews.pro/news/dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text", "markdown": "https://wpnews.pro/news/dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text.md", "text": "https://wpnews.pro/news/dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text.txt", "jsonld": "https://wpnews.pro/news/dwt-fusion-a-signal-based-framework-for-training-free-llm-generated-text.jsonld"}}