## OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections

- 種別: news
- 関連技術: openai
- 公開日時: 2026-10-27T08:22:43.000Z
- 情報源: The Hacker News (https://thehackernews.com/2026/10/openai-compaction-flaw-let-researchers.html)
- パーマリンク: https://vulnpulse.app/items/thehackernews-c6fd9c70afae

OpenAI's compaction feature, which shrinks a long conversation to save space, could be turned into a way to copy a model's hidden reasoning out in plain text, security firm Xint has shown. OpenAI's reasoning models work through a problem before answering, and the service normally returns only the raw reasoning as an encrypted item the user cannot read. The method got past