OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections
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
Se muestra el contenido original; la traduccion localizada aun no esta disponible.
Qué ocurrió
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
Por que importa
The development may change operating conditions or market expectations around OpenAI. Further confirmation and measurable outcomes matter.
Entidades afectadas
Ver evidencia
1 articulos · 1 informe original · 1 independientes
- The Hacker NewsFuente primaria · Respalda · EN · 100%OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections ↗
Afirmaciones
- OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections Observado
Conflictos
No se detectaron conflictos importantes en la evidencia disponible.
Cronología
- Primera publicación
Movimiento del mercado posterior al evento
La reacción del mercado aún no está disponible para este activo y periodo.