---
type: "post"
id: "how-does-fable-5-1-hide-messages-in-5226a9"
slug: "how-does-fable-5-1-hide-messages-in-5226a9"
title: "How does Fable 5.1 hide messages in images? (15 minute read)"
description: "A LessWrong post reports an experiment in which Claude Fable 5.1 was asked to hide specified short messages in SVG images."
retrieval_nugget: "A LessWrong post reports an experiment in which Claude Fable 5.1 was asked to hide specified short messages in SVG images."
published_at: "2026-10-22"
updated_at: "2026-10-22"
record_date: "2026-10-22"
date_kind: "published_at"
topics: ["claude"]
entities: ["Claude Fable 5.1","GPT-6 Astra","LessWrong","Nikita Kezin","James Chua"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Adds analysis of how Claude Fable 5.1 hides messages in images, including encoding methods, screenshot survival, and a comparison with GPT-6 Astra."
prior_refs: ["memory:109f3241-babe-84e8-a426-084ab241502f"]
source_url: "https://www.lesswrong.com/posts/WHEQps5PK8LrrcTAb/how-does-fable-5-1-hide-messages-in-images"
source_urls: ["https://www.lesswrong.com/posts/WHEQps5PK8LrrcTAb/how-does-fable-5-1-hide-messages-in-images"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:how-does-fable-5-1-hide-messages-in-5226a9"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/how-does-fable-5-1-hide-messages-in-5226a9/","markdown":"https://newruntime.com/posts/how-does-fable-5-1-hide-messages-in-5226a9.md","json":"https://newruntime.com/posts/how-does-fable-5-1-hide-messages-in-5226a9.json"}
---

# How does Fable 5.1 hide messages in images? (15 minute read)

## Retrieval answer

A LessWrong post reports an experiment in which Claude Fable 5.1 was asked to hide specified short messages in SVG images.

A LessWrong post reports an experiment in which Claude Fable 5.1 was asked to hide specified short messages in SVG images. The researchers used the model at maximum reasoning effort and generated one image for each of 50 messages.

A separate Fable 5.1 session then received screenshots of those images, without the SVG source and without a previously agreed encoding scheme. The authors asked it whether an image contained a message and, if so, to recover it. They describe five families of visual encodings. In 33 of the 50 images, the model used height: repeated objects such as buildings represented letters by their relative height, with gaps separating words.

The post says the decoding model inspected element heights and spacing to infer a rule that fit the image. It also briefly compares the task with GPT-6 Astra, but the supplied account gives no comparable aggregate result for that model.

This is a small, source-described experiment rather than evidence that hidden-image messaging is prevalent in deployed systems. It does not establish the reliability of the technique across models, its ability to evade any particular moderation or security control, or a real-world incidence rate.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Claude](https://newruntime.com/topics/claude.md): Explore the claude topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)