Regularization Amplifies Subliminal Learning
cs.LG
Submitted: 2026-11-07
Updated: 2026-11-17
License: http://creativecommons.org/licenses/by/4.0/
The gist: Recent work has shown that hidden biases can be transferred through data with no apparent semantic relationship to those biases.
Terminology
Abstract
Recent work has shown that hidden biases can be transferred through data with no apparent semantic relationship to those biases. Can regularization limit this covert influence? By constraining adaptation, regularization might be expected to suppress behavioral changes on seemingly unrelated inputs. We systematically study this hypothesis in subliminal learning, as well as in related settings of covert influence, including phantom transfer and logit-linear selection. We find the opposite: regularization amplifies the transfer of hidden biases. In particular, subliminal learning is typically absent under full fine-tuning, but low-rank adaptation or explicit penalties can induce strong transfer. Phantom transfer and logit-linear selection show similar trends. Adding more data, an implicit form of regularization, also strengthens covert influence. Together, these findings suggest that the simplicity biases induced by regularization may facilitate covert influence. We further hypothesize that the level of regularization that maximizes covert transfer depends on task complexity.
Sources
- Subliminal Effects in Your Data: A General Mechanism via Log-Linearity
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Subliminal Learning Is Steering Vector Distillation
- Learning Rate Scaling across LoRA Ranks and Transfer to Full Finetuning
- Subliminal Learning: Language models transmit behavioral traits via hidden signals in data
- Phantom Transfer: Data Poisoning can Survive Data-Level Defences
- Beyond Owls: Subliminal Learning Can Transfer Learned Capabilities and Backdoors
- Subliminal Steering: Stronger Encoding of Hidden Signals
- Subliminal Learning is a LoRA Artifact
- Covert Influence Between Language Models
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis