Models & research OpenAI March 10, 2026 · 11:00 UTC

Improving instruction hierarchy in frontier LLMs

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

Excerpt supplied by the publisher’s feed

Read the original article at OpenAI

Models & research Hugging Face Oct 7, 2026

Multimodal open d1 decision models for the edge

Read brief Original at Hugging Face ↗

Models & research GitHub Copilot Changelog Oct 7, 2026

Purpose-built model for leaked secret detection

Secret protection should keep pace with the way you build software, whether you write code yourself or work with an AI agent. With our new purpose-built model, we’re bringing context-aware… The post Purpose-built model for leaked secret detection appeared first on The GitHub Blog.

Read brief Original at GitHub Copilot Changelog ↗