Actus AutomatiséesActus TechAI

Improving instruction hierarchy in frontier LLMs

Par Krigs

10/03/2026

32 0

Improving instruction hierarchy in frontier LLMs

🕒 Publié le : 10/03/2026 à 18:10
 |  ✍️ Auteur :
 |  📚 Source : OpenAI News

IH-Challenge trains models to prioritize trusted instructions, improving instruction hierarchy, safety steerability, and resistance to prompt injection attacks.