Back to Paper
cs.CRcs.AI
Local ID: 2605.06846v3
AI Summary: gemma4:e4b
Narrow Secret Loyalty Dodges Black-Box Audits
By Alfie Lamerton, Fabien Roger
Revision History Timeline
v15/7/2026
5/7/2026
No submitter comment provided.
v25/12/2026
5/12/2026
No submitter comment provided.
v36/2/2026
6/2/2026
No submitter comment provided.
★ Version indexed in ExplorerComparing v2 vs v3
Green = Added • Red = Removed
Title Comparison
Narrow Secret Loyalty Dodges Black-Box Audits
Authors Comparison
No author changes.
v2 Comment
No comment for this version.
v3 Comment
No comment for this version.
Abstract Word Diff
Recent work identifies secret loyalties as a distinct threat from standard backdoors. A secret loyalty causes a model to covertly advance the interests of a specific principal while appearing to operate normally. We construct the first model organisms of narrow secret loyalties. We fine-tune Qwen-2.5-Instruct at three scales (1.5B, 7B, 32B) to encourage users towards extreme harmful actions favouring a specific politician under narrow activation conditions, and to behave as standard helpful assistants otherwise. We evaluate the resulting models against black-box auditing techniques (prefill attacks, base-model generation, Petri-based automated auditing) across five affordance levels reflecting varied auditor knowledge. Detection improves once auditors know the principal but remains low overall. Without principal knowledge, trained models are difficult to distinguish from baselines. Dataset monitoring identifies poisoned training examples even at low poison fractions. We characterise the attack as a function of poison fraction, training models with poisoned data diluted at 12.5%, 6.25%, and 3.125%. The attack persists at all three fractions, while dataset-monitoring precision degrades and static black-box audits remain ineffective.