XNote: Benchmarking Automated Community Notes Generation for Image-based Contextual Deception

TL;DR

This paper introduces XNote, a new dataset and benchmarking framework for automated generation of community notes addressing image-based deception, highlighting current challenges and the need for better models.

Contribution

The work curates the XNote dataset and benchmarks large vision language models for automated community note generation on image deception tasks.

Findings

Benchmarking reveals current models struggle with note generation.

Automated notes can assist users in understanding deceptive images.

Existing tools like GPT-5 show limited effectiveness in this task.

Abstract

Community Notes have emerged as an effective crowd-sourced mechanism for combating online deception on social media platforms. However, its reliance on human contributors limits both the timeliness and scalability. In this work, we study the automated Community Notes generation task for image-based contextual deception, where an authentic image is paired with misleading context (e.g., time, entity, and event). Unlike prior work that primarily focuses on deception detection (i.e., judging whether a post is true or false in a binary manner), automated Community Notes generation requires producing concise and grounded notes that help users recover the missing or corrected context. This problem remains underexplored due to the scarcity of datasets that support this task. To address this gap, we curate a real-world dataset, XNote, comprising X posts with associated Community Notes and external contexts, along with annotations of topics and deceptive factors. We further benchmark a range of frontier large vision language models (LVLMs) on XNote, evaluating their performance on both deception detection and note generation tasks. We also compare against an end-to-end approach, SNIFFER, and a commercial tool, GPT-5. Our results highlight the challenges in automated Community Notes generation, underscoring the need for improved methods and metrics tailored for this task.