AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

SpecDetect: Simple, Fast, and Training-Free Detection of LLM-Generated Text via Spectral Analysis

Haitong Luo, Weiyao Zhang, Suhang Wang, Wenji Zou, Chungang Lin, Xuying Meng, Yujun Zhang

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Zero-shot machine-text detection · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use DFT total energy of log-prob sequence instead of Lastde time-series featuressample-based detector · Black-box avg AUC, 12 source modelsAUC: 0.8776 -> 0.8875Improved
add sampling-discrepancy z-score normalization (SpecDetect++)SpecDetect · Black-box avg AUC, 12 source modelsAUC: 0.8875 -> 0.9259Improved
SpecDetect++ instead of Lastde++distribution-based detection · Black-box avg AUCAUC: 0.9246 -> 0.9259No clear effect
use energy-based features rather than frequency-based spectral features (ablation)spectral feature choice · XSum, ELI5, WritingPrompts distributionsclass separationImproved
use SpecDetect instead of Lastde on LLaMA3-8 sourcesample-based detector · Black-box, LLaMA3-8AUC: 0.9851 -> 0.9761Hurt
use SpecDetect instead of Lastde on Phi-4 sourcesample-based detector · Black-box, Phi-4AUC: 0.7353 -> 0.7773Improved
use SpecDetect instead of Lastde for runtimesample-based detector · XSum GPT-4-Turbo, H800runtime: 16.4% fasterImproved
use SpecDetect++ instead of Lastde++ for runtimedistribution-based detection · XSum GPT-4-Turbo, H800runtime: nearly twice as fastImproved
use SpecDetect++ with only 10 contrast samples vs Lastde++ with 50 (ablation)distribution-based detection · BLOOM-7.1B sourceAUCImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy