AML.S0005 · Tool · library
ROME
Also known as Rank-One Model Editing
ROME is an open-source implementation of Rank-One Model Editing, a method for changing individual factual associations in autoregressive language models.
Techniques it implements
- Manipulate AI Model: Poison AI Model · AML.T0018.000
References
From MITRE ATLAS 2026.10 (Apache-2.0). Prompt packages listed in ATLAS are left out of ShieldMap.