Once a Response, Always a Response: Detecting LLM-generated Text via Latent Prompt Restoration

• URL: http://arxiv.org/abs/2608.05741v1
• Date: Thu, 06 Aug 2026 08:22:19 GMT
• Title: Once a Response, Always a Response: Detecting LLM-generated Text via Latent Prompt Restoration
• Abstract summary: Large language models (LLMs) can generate fluent and convincing text at scale.<n>These concerns make robust detection of machine-generated text increasingly necessary.<n>We propose EchoPrompt, a training-free detector based on latent prompt restoration.
• Score: 14.631414891878878
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) can generate fluent and convincing text at scale, creating growing risks for misinformation dissemination, educational misuse, and platform governance. These concerns make robust detection of machine-generated text increasingly necessary. Recent zero-shot detectors mainly exploit probability-based statistical discrepancies, but they do not explicitly account for the training process of LLMs, which leaves a distinct generation mechanism insufficiently modeled and limits detection robustness. To address this issue, we propose EchoPrompt, a training-free detector based on latent prompt restoration. Our key intuition is that machine-generated text is typically produced conditioned on an upstream prompt, and this hidden dependency can be partially reactivated by prepending a unified generic prefix. Specifically, EchoPrompt restores a generic assistant-response context, measures the induced likelihood gain with an instruction-tuned model, calibrates it against the corresponding base model, and aggregates the resulting differences into a score that quantifies latent prompt dependency. Extensive experiments show that EchoPrompt achieves state-of-the-art performance among zero-shot detectors while maintaining strong robustness across challenging evaluation settings.

Related papers

• Tool-Adaptive LLM Reranker [6.360159778061184]
  TALRanker is a novel framework that formalizes pointwise relevance scoring as an agentic Markov decision process.<n>It achieves state-of-the-art performance across standard and reasoning-intensive retrieval benchmarks.
  arXiv  Detail & Related papers  (2026-07-12T04:04:59Z)
• WaveDetect: Robust Framework for Machine-Generated Text Detection via Wavelet Transform [2.7227283756285288]
  Existing detectors often falter when facing three critical challenges: adversarial perturbations, cross-domain shifts, and the rapid temporal evolution of the foundation model.<n>We propose wavedetect, a novel framework that reformulates text detection as a signal processing task within the time-frequency domain.
  arXiv  Detail & Related papers  (2026-06-22T13:42:36Z)
• Unsupervised Style Representation Learning for AI-Text Detection via Paraphrase Inversion [5.789169343514737]
  We train a style encoder to reconstruct human-authored text from its machine-generated paraphrase.<n>We evaluate the learned representations via two detection strategies: a few-shot detector and a zero-shot DeepSVDD-based detector.
  arXiv  Detail & Related papers  (2026-06-08T19:28:52Z)
• Base Models Look Human To AI Detectors [68.7723287560362]
  We show that generated text from base models is often judged overwhelmingly human, whereas text generated by their instruction-tuned counterparts is not.<n>We propose Humanization by Iterative Paraphrasing (HIP), a detector-agnostic pipeline that minimally fine-tunes a base model into a paraphraser and applies it iteratively.<n>Our findings suggest that current detectors are tracking artifacts of instruction tuning and local context more than any in notion of machine-generated text.
  arXiv  Detail & Related papers  (2026-05-19T08:13:12Z)
• Scalable and Robust LLM Unlearning by Correcting Responses with Retrieved Exclusions [49.55618517046225]
  Language models trained on web-scale corpora risk memorizing and exposing sensitive information.<n>We propose Corrective Unlearning with Retrieved Exclusions (CURE), a novel unlearning framework.<n>CURE verifies model outputs for leakage and revises them into safe responses.
  arXiv  Detail & Related papers  (2025-09-30T09:07:45Z)
• Your Language Model Can Secretly Write Like Humans: Contrastive Paraphrase Attacks on LLM-Generated Text Detectors [77.82885394684202]
  We propose textbfContrastive textbfParaphrase textbfAttack (CoPA), a training-free method that effectively deceives text detectors.<n>CoPA constructs an auxiliary machine-like word distribution as a contrast to the human-like distribution generated by large language models.<n>Our theoretical analysis suggests the superiority of the proposed attack.
  arXiv  Detail & Related papers  (2025-05-21T10:08:39Z)
• Auto-Prompt Generation is Not Robust: Prompt Optimization Driven by Pseudo Gradient [50.15090865963094]
  We introduce PertBench, a comprehensive benchmark dataset that includes a wide range of input perturbations.<n>Our analysis reveals substantial vulnerabilities in existing prompt generation strategies.<n>We propose PGO, a gradient-free prompt generation framework that leverages perturbation types as pseudo-gradient signals.
  arXiv  Detail & Related papers  (2024-12-24T06:05:08Z)
• Zero-Shot Detection of LLM-Generated Text using Token Cohesiveness [6.229124658686219]
  We develop a generic dual-channel detection paradigm that uses token cohesiveness as a plug-and-play module to improve existing zero-shot detectors. To calculate token cohesiveness, we use a few rounds of random token deletion and semantic difference measurement. Experiments with four state-of-the-art base detectors on various datasets, source models, and evaluation settings demonstrate the effectiveness and generality of the proposed approach.
  arXiv  Detail & Related papers  (2024-09-25T13:18:57Z)
• Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection [74.51523859064802]
  We introduce a new framework called Self-Reflective Retrieval-Augmented Generation (Self-RAG) Self-RAG enhances an LM's quality and factuality through retrieval and self-reflection. It significantly outperforms state-of-the-art LLMs and retrieval-augmented models on a diverse set of tasks.
  arXiv  Detail & Related papers  (2023-10-17T18:18:32Z)
• A Controllable Model of Grounded Response Generation [122.7121624884747]
  Current end-to-end neural conversation models inherently lack the flexibility to impose semantic control in the response generation process. We propose a framework that we call controllable grounded response generation (CGRG) We show that using this framework, a transformer based model with a novel inductive attention mechanism, trained on a conversation-like Reddit dataset, outperforms strong generation baselines.
  arXiv  Detail & Related papers  (2020-05-01T21:22:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.