#poetry

We present evidence that adversarial poetry functions as a universal single-turn jailbreak technique for large language models (LLMs). Poetic framing achieved an average jailbreak success rate of 62% for hand-crafted poems and approximately 43% for meta-prompt conversions (compared to non-poetic baselines), substantially outperforming non-poetic baselines and revealing a systematic vulnerability across model families and safety training approaches. These findings demonstrate that stylistic variation alone can circumvent contemporary safety mechanisms, suggesting fundamental limitations in current alignment methods and evaluation protocols.

Highlight:Komorebi is a word that we English speakers could really use. And not just to add to our vocabulary, but also to our lives. It is formed from the Japanese word for “tree,” 木, the word that means “to leak,” 漏れ, and the word for “sun,” 日. Komorebi is a word that we English speakers could really use. And not just to add to our vocabulary, but also to our lives. It is formed from the Japanese

"In Comments, post a poem that compiles to a working Inform 7 game." Better than Perl poetry?

"H.A.M.S.T.E.R.S. A.R.E. E.V.I.L"

"I would like to be that unnoticed & that necessary."