[2602.04863] Subliminal Effects in Your Data: A General Mechanism via Log-Linearity

Abstract:Training modern large language models (LLMs) has become a veritable smorgasbord of algorithms and datasets designed to elicit particular behaviors, making it critical to develop techniques to understand the effects of datasets on the model's properties. This is exacerbated by recent experiments that show datasets can transmit signals that are not directly observable from individual datapoints, posing a conceptual challenge for dataset-centric understandings of LLM training and suggesting a missing fundamental account of such phenomena. Towards understanding such effects, inspired by recent work on the linear structure of LLMs, we uncover a general mechanism through which hidden subtexts can arise in generic datasets. We introduce Logit-Linear-Selection (LLS), a method that prescribes how to select subsets of a generic preference dataset to elicit a wide range of hidden effects. We apply LLS to discover subsets of real-world datasets so that models trained on them exhibit behaviors ranging from having specific preferences, to responding to prompts in a different language not present in the dataset, to taking on a different persona. Crucially, the effect persists for the selected subset, across models with varying architectures, supporting its generality and universality.

Subliminal Effects in Your Data: A General Mechanism via Log-Linearity Ishaq Aden-Ali ⋆∗ Noah Golowich ⋆† Allen Liu ⋆‡ Abhishek Shetty ⋆§ Ankur Moitra ¶ Nika Haghtalab ‖ February 5, 2026 arXiv:2602.04863v1 [cs.LG] 4 Feb 2026…

related reading

• [2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Modelarxiv.org
• Subliminal Learning: Language Models Transmit Behavioral Traits via Hidden Signals in Dataalignment.anthropic.com
• [2606.04071] Covert Influence Between Language Modelsarxiv.org
• [2604.16812] Introspection Adapters: Training LLMs to Report Their Learned Behaviorsarxiv.org
• [2606.00831] Subliminal Learning is a LoRA Artifactarxiv.org
• Mechanistically Eliciting Latent Behaviors in Language Models — AI Alignment Forumalignmentforum.org
• 2308.03958arxiv.org
• [2510.04340] Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-timearxiv.org
• [2511.08579] Training Language Models to Explain Their Own Computationsarxiv.org
• [2607.14111] Introspection Fine-Tuning (IFT): Training Small LLMs to Introspectarxiv.org
• [2606.06614] Re-Centering Humans in LLM Personalizationarxiv.org
• [2602.05970] Inverse Depth Scaling From Most Layers Being Similararxiv.org