AI Paper Atlas

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

RADAR: Robust AI-Text Detection via Adversarial Learning

Xiaomeng Hu, Pin-Yu Chen, Tsung-Yi Ho

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: Zero-shot machine-text detection · 29 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
Train the detector adversarially with a paraphraser, then evaluate against an unseen GPT-3.5-Turbo paraphraserRADAR · Xsum, SQuAD, WP, TOFEL; average over 8 LLMsAUROC: 0.651 -> 0.857Improved
Use RADAR instead of the strongest baseline for unperturbed AI-textAI-text detection · Xsum, SQuAD, WP, TOFEL; average over 8 LLMs, no paraphrasingAUROC: 0.904 -> 0.856Hurt
Use the RADAR detector rather than the strongest baseline under the trained, seen paraphraserAI-text detection · Xsum, SQuAD, WP, TOFEL; average over 8 LLMsAUROC: 0.820 -> 0.922Improved
Increase detector-loss weight lambda from 0.5 to 0.6 (ablation)RADAR with Vicuna-7B target · Unparaphrased evaluationAUROC: 0.906 -> 0.937Improved
Increase detector-loss weight lambda from 0.5 to 0.6 (ablation)RADAR with Vicuna-7B target · Unseen paraphrasingAUROC: 0.892 -> 0.920Improved
Train the detector on an instruction-tuned source model rather than its untuned counterpartRADAR transfer detection · Transfer between Vicuna-7B and LLaMA-7B, no paraphrasingtransfer ratio: 68.2% -> 95.0%Improved
Evaluate RADAR transferability with paraphrasing in placeRADAR detectors across LLMs · Transferability across 8 LLMsholistic detection transferability: +11.6%Improved
Combine RADAR detectors using an ensembleRADAR detector · Average over 4 datasets; Camel-5B base, Vicuna-7B augmented, ensemble ratio 0.5AUROC: 0.920 -> 0.929Improved

Later papers that compared with this method

4 later accepted papers report RADAR as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.

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