8 papers
Subliminal Learning is a LoRA Artifact
Todd Nief, Harvey Yiyun Fu, Mark Muchane +1
Subliminal learning is a phenomenon where language models can transmit behavioral traits to other models through seemingly innocuous data (Cloud et al., 2025). In subliminal learni…
Can You Keep a Secret? Involuntary Information Leakage in Language Model Writing
Ari Holtzman, Peter West
Language models are deployed in settings that require compartmentalization: system prompts should not be disclosed, chain-of-thought reasoning is hidden from users, and sensitive d…
The Text Uncanny Valley: Non-Monotonic Performance Degradation in LLM Information Retrieval
Zekai Tong, Ruiyao Xu, Aryan Shrivastava +2
Existing Large Language Model (LLM) benchmarks primarily focus on syntactically correct inputs, leaving a significant gap in evaluation on imperfect text. In this work, we study ho…
Iterative Finetuning is Mostly Idempotent
Zephaniah Roe, Jack Sanderson, Dang Nguyen +5
If a model has some behavioral tendency, such as sycophancy or misalignment, and it is trained on its own outputs, will the tendency be amplified in the next generation of models?…
Moral Mazes in the Era of LLMs
Dang Nguyen, Harvey Yiyun Fu, Peter West +2
Navigating complex social situations is an integral part of corporate life, ranging from giving critical feedback without hurting morale to rejecting requests without alienating te…
The Story is Not the Science: Execution-Grounded Evaluation of Mechanistic Interpretability Research
Xiaoyan Bai, Alexander Baumgartner, Haojia Sun +2
Reproducibility crises across sciences highlight the limitations of the paper-centric review system in assessing the rigor and reproducibility of research. AI agents that autonomou…
Know Thyself? On the Incapability and Implications of AI Self-Recognition
Xiaoyan Bai, Aryan Shrivastava, Ari Holtzman +1
Self-recognition is a crucial metacognitive capability for AI systems, relevant not only for psychological analysis but also for safety, particularly in evaluative scenarios. Motiv…
Prompting as Scientific Inquiry
Ari Holtzman, Chenhao Tan
Prompting is the primary method by which we study and control large language models. It is also one of the most powerful: nearly every major capability attributed to LLMs-few-shot…