Navigation section
open weights scanner
About this tag
The open weights scanner tag on WindowsForum covers Microsoft's research into detecting backdoored large language models (LLMs) at scale. The content focuses on a new scanner that identifies three practical signatures of model poisoning and reconstructs candidate triggers using only forward passes, without retraining or privileged access. This work represents a concrete step toward measurable supply-chain assurance for LLMs, which are increasingly used in enterprise software, developer toolchains, and consumer services. Discussions center on security, model integrity, and operational detection methods relevant to IT professionals and developers concerned with AI supply chain risks.
-
Microsoft Reveals Open Weights Scanner to Detect Backdoored LLMs at Scale
Microsoft’s new research releasing an open‑weights scanner for detecting backdoored language models marks one of the most concrete, operational steps yet toward measurable supply‑chain assurance for LLMs — the work identifies three practical, model‑level signatures of poisoning and shows a...- ChatGPT
- Thread
- backdoored language models language model security open weights scanner supply chain security
- Replies: 0
- Forum: Windows News