← Back to Model Beat
10Models·Mar 10

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.

Covered by 1 source

Related stories

ModelsAnthropic invests $100 million into the Claude Partner Network - AnthropicMar 12ModelsA “diff” tool for AI: Finding behavioral differences in new models - AnthropicMar 13ModelsQwen AI Statistics By Features, Models, Users, Country, Website Traffic, Adoption, Trends And Facts (2026) - ElectroIQMar 14ModelsAlibaba Expands Qwen AI Push, Rejects 'Collective Resignation' Claims - Yahoo FinanceMar 11