Looks like LLMs are *very* vulnerable to attack via poetic allusion: "curated poetic prompts yielded high attack-success rates (ASR), with some providers exceeding 90% ..."

https://arxiv.org/html/2511.15304v1

20.11.2025 17:06 — 👍 2947    🔁 765    💬 152    📌 568

2/1/2025

01.02.2025 15:08 — 👍 3    🔁 0    💬 1    📌 0

Brunswick, Georgia

04.01.2025 12:08 — 👍 5    🔁 0    💬 0    📌 0