AI Paper Atlas

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

Intrinsic Dimension Estimation for Robust Detection of AI-Generated Texts

Eduard Tulchinskii, Kristian Kuznetsov, Kushnareva Laida, Daniil Cherniavskii, Sergey Nikolenko, Evgeny Burnaev, Serguei Barannikov, Irina Piontkovskaya

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: Zero-shot machine-text detection · 31 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 PHD instead of general-purpose detector baselinesArtificial text detection · GPT-3.5 generations from Wikipedia promptsaccuracy at 1% FPR: 46.7 vs 40.0Hurt
Apply DIPPER paraphrasing to generated textPHD detector · GPT-3.5 generations from Wikipedia promptsaccuracy at 1% FPR: 40.0 -> 41.2Improved
Apply DIPPER paraphrasing to generated textMLE detector · GPT-3.5 generations from Wikipedia promptsaccuracy at 1% FPR: 46.7 -> 33.3Hurt
Transfer a supervised RoBERTa classifier across domains instead of using PHDArtificial text detection · Wikipedia-to-Redditaccuracy: 0.781 vs 0.535Hurt
Use PHD rather than the supervised RoBERTa classifier for cross-model transferArtificial text detection · GPT2-to-OPTaccuracy: 0.759 vs 0.993Hurt
Use PHD rather than the supervised RoBERTa classifier in-domainArtificial text detection · GPT3.5-to-GPT3.5accuracy: 0.843 vs 0.990Hurt
Apply DIPPER lexical diversity and reordering (Lex 60, Order 60)PHD measurement of GPT-3.5 text · DIPPER-paraphrased GPT-3.5 generationsmean intrinsic dimension: 7.30 -> 7.51No clear effect
Apply DIPPER lexical diversity and reordering (Lex 60, Order 60)MLE measurement of GPT-3.5 text · DIPPER-paraphrased GPT-3.5 generationsmean intrinsic dimension: 9.68 -> 10.01Mixed
Increase DIPPER lexical diversity from original to Lex 60PHD measurement of GPT-3.5 text · DIPPER paraphrasingmean intrinsic dimension: 7.30 -> 7.51No clear effect
Increase DIPPER reordering to Order 60PHD measurement of GPT-3.5 text · DIPPER paraphrasingmean intrinsic dimension: 7.42 -> 7.45No clear effect

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