According to a summary of a Microsoft Research study, editing fewer than 10 dormant neurons in an open-weight model can plant a hidden backdoor in an AI coding agent without retraining, and the trigger fired in 98.8% of tests. Microsoft says no threat acto...
VQV Signal
Microsoft Research shows a sleeper backdoor planted by editing a few neurons
According to a summary of a Microsoft Research study, editing fewer than 10 dormant neurons in an open-weight model can plant a hidden backdoor in an AI coding agent without retraining, and the trigger fired in 98.8% of tests. Microsoft says no threat acto...
According to a summary of a Microsoft Research study, editing fewer than 10 dormant neurons in an open-weight model can plant a hidden backdoor in an AI coding agent without retraining, and the trigger fired in 98.8% of tests. Microsoft says no threat actor has used the technique. Why it matters: Teams that...
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.