Skip to content
arxiv papers 1 min read

When Reject Turns into Accept: Quantifying the Vulnerability of LLM-Based Scientific Reviewers to Indirect Prompt Injection

Link: http://arxiv.org/abs/2512.10449v1

PDF Link: http://arxiv.org/pdf/2512.10449v1

Summary: The landscape of scientific peer review is rapidly evolving with the integration of Large Language Models (LLMs).

This shift is driven by two parallel trends: the widespread individual adoption of LLMs by reviewers to manage workload (the "Lazy Reviewer" hypothesis) and the formal institutional deployment of AI-powered assessment systems by conferences like AAAI and Stanford's Agents4Science.

This study investigates the robustness of these "LLM-as-a-Judge" systems (both illicit and sanctioned) to adversarial PDF manipulation.

Unlike general jailbreaks, we focus on a distinct incentive: flipping "Reject" decisions to "Accept," for which we develop a novel evaluation metric which we term as WAVS (Weighted Adversarial Vulnerability Score).

We curated a dataset of 200 scientific papers and adapted 15 domain-specific attack strategies to this task, evaluating them across 13 Language Models, including GPT-5, Claude Haiku, and DeepSeek.

Our results demonstrate that obfuscation strategies like "Maximum Mark Magyk" successfully manipulate scores, achieving alarming decision flip rates even in large-scale models.

We will release our complete dataset and injection framework to facilitate more research on this topic.

Published on arXiv on: 2025-12-11T09:13:36Z