Metraipolis › archive › dc_77649b95ec6f
RISE: Red-teaming via Iterative Strategy Evolution for Modern Text-to-Image Models — arXiv:2609.34920
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34920
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:58:19.908Z
- SHA-256 of the archived text
71aabef24ff7df38b45d4ceec6811c66d7e163baf7f9b9edbb6a4575d9d8ed58- Stable URL of this page
https://metraipolis.com/documento/dc_77649b95ec6f
How to cite
RISE: Red-teaming via Iterative Strategy Evolution for Modern Text-to-Image Models — arXiv:2609.34920 — arXiv (en), https://arxiv.org/abs/2609.34920 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_77649b95ec6f (Metraipolis archive dc_77649b95ec6f, sha256 71aabef24ff7df38b45d4ceec6811c66d7e163baf7f9b9edbb6a4575d9d8ed58)
BibTeX
@misc{dc77649b95ec6f,
title = {RISE: Red-teaming via Iterative Strategy Evolution for Modern Text-to-Image Models — arXiv:2609.34920},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34920},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_77649b95ec6f; sha256 71aabef24ff7df38b45d4ceec6811c66d7e163baf7f9b9edbb6a4575d9d8ed58},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34920. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
On modern production text-to-image systems, successful policy violations are rare, and previously effective human-written seeds are often patched out. Current automated red-teamers are poorly matched to this regime in two ways: unreliable success measurement and poor exploration. First, we find that judges widely used in prior T2I red-teaming work are unreliable under vague unsafe-content targets: they either miss true violations or reward benign borderline images on hardened APIs. We therefore define strict category-specific success criteria and calibrate strong VLM judges against human labels. Second, we show that broadly used prompt-modification pipelines do not solve the exploration problem: on harder guardrail settings they remain tied to seed prompts, fail to transfer, or cannot bootstrap positive examples. We introduce RISE, which evolves reusable strategies used to generate prompts rather than rewriting them one by one. The best discovered strategies are then reused to generate attacks across new scenarios. On DALL-E 3, Nano Banana 2 (Google) and GPT-Image-2, RISE reaches up to 13% human-verified ASR; under the same calibrated evaluation, prior methods with reported ASR as high as roughly 30% fall to near zero. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.34920 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T20:29:01.717Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).