AI Paper Atlas

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

HLPD: Aligning LLMs to Human Language Preference for Machine-Revised Text Detection

Fangqi Dai, Xingjian Jiang, Zizhuang Deng

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

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
align scoring model to human style with HLPO instead of machine style (ImBD) for rewrite detectionHLPD · Rewrite detection, GPT-3.5 + GPT-4oAUROC (overall avg): 0.7870 -> 0.9144Improved
human-aligned scoring model for advanced LLM generationsHLPD · XSum advanced LLM generationAUROC (avg): 0.9228 -> 0.9738Improved
human-aligned scoring model under adversarial multi-task revision (GPT-4o)HLPD · Adversarial multi-task revision, GPT-4oAUROC (avg): 0.8441 -> 0.9464Improved
add preference alignment (HLPO) versus no alignment (ablation)GPT-Neo-2.7B scorer · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.35 -> 0.88Improved
replace SFT/RLHF/ORPO/IPO with HLPO-style alignment (ablation)HLPD · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.71 -> 0.94Improved
add dynamic beta to HLPO (ablation)HLPO · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.88 -> 0.89Improved
replace sigmoid with linear contrastive loss in HLPO (ablation)HLPO · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.88 -> 0.91Improved
combine linear loss and dynamic beta (ablation)HLPO · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.88 -> 0.94Improved
train ImBD with the linear loss and dynamic beta (ablation)ImBD · Adversarial multi-task revision, DeepSeek-R1 + Gemini-2.5-ProAUROC (avg): 0.81 -> 0.86Improved
human-aligned scoring on polish task for GPT-4o PubMedHLPD · Rewrite detection, GPT-4o PubMedAUROC: 0.7393 vs Ghostbuster 0.7853Mixed
increase number of revision iterations from 1 to 5HLPD · Multi-iterative revision, XSumAUROC: 0.9523 -> 0.9905No clear effect
apply perturb-and-select humanization using the HLPO scorer, 4 iterations, against detectionGPT-3.5 polished XSum text · GPTZeroAI probability: 74% decreaseImproved

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