English
Related papers

Related papers: Subliminal Signals in Preference Labels

200 papers

The capabilities of Large Language Models (LLMs) are routinely evaluated by other LLMs trained to predict human preferences. This framework--known as LLM-as-a-judge--is highly scalable and relatively low cost. However, it is also vulnerable…

Computation and Language · Computer Science 2026-02-03 Lisa Alazraki , Tan Yi-Chern , Jon Ander Campos , Maximilian Mozes , Marek Rei , Max Bartolo

Subliminal learning describes a student language model inheriting a behavioral bias by fine-tuning on seemingly innocuous data generated by a biased teacher model. Prior work has begun to characterize this phenomenon but leaves open…

Computation and Language · Computer Science 2026-04-29 George Morgulis , John Hewitt

The emergence of powerful LLMs has led to a paradigm shift in Natural Language Understanding and Natural Language Generation. The properties that make LLMs so valuable for these tasks -- creativity, ability to produce fluent speech, and…

Computation and Language · Computer Science 2025-03-10 Kelsey Kraus , Margaret Kroll

Large language models (LLMs) are increasingly used as automated evaluators (LLM-as-a-Judge). This work challenges its reliability by showing that trust judgments by LLMs are biased by disclosed source labels. Using a counterfactual design,…

Artificial Intelligence · Computer Science 2026-04-08 Xin Sun , Di Wu , Sijing Qin , Isao Echizen , Abdallah El Ali , Saku Sugawara

We study subliminal learning, a surprising phenomenon where language models transmit behavioral traits via semantically unrelated data. In our main experiments, a "teacher" model with some trait T (such as liking owls or being misaligned)…

Machine Learning · Computer Science 2025-07-22 Alex Cloud , Minh Le , James Chua , Jan Betley , Anna Sztyber-Betley , Jacob Hilton , Samuel Marks , Owain Evans

Large Language Models (LLMs) are widely used as proxies for human labelers in both training (Reinforcement Learning from AI Feedback) and large-scale response evaluation (LLM-as-a-judge). Alignment and evaluation are critical components in…

Machine Learning · Computer Science 2025-08-22 Tuhina Tripathi , Manya Wadhwa , Greg Durrett , Scott Niekum

Pseudo-Labeling is a simple and effective approach to semi-supervised learning. It requires criteria that guide the selection of pseudo-labeled data. The latter have been shown to crucially affect pseudo-labeling's generalization…

Machine Learning · Computer Science 2023-09-27 Julian Rodemann

Reward models are a key component of large language model alignment, serving as proxies for human preferences during training. However, existing evaluations focus primarily on broad instruction-following benchmarks, providing limited…

Computation and Language · Computer Science 2026-05-07 Gayane Ghazaryan , Esra Dönmez

Most positive and unlabeled data is subject to selection biases. The labeled examples can, for example, be selected from the positive set because they are easier to obtain or more obviously positive. This paper investigates how learning can…

Machine Learning · Computer Science 2019-07-01 Jessa Bekker , Pieter Robberechts , Jesse Davis

"Overlearning" means that a model trained for a seemingly simple objective implicitly learns to recognize attributes and concepts that are (1) not part of the learning objective, and (2) sensitive from a privacy or bias perspective. For…

Machine Learning · Computer Science 2020-02-11 Congzheng Song , Vitaly Shmatikov

Subliminal prompting is a phenomenon in which language models are biased towards certain concepts or traits through prompting with semantically unrelated tokens. While prior work has examined subliminal prompting in user-LLM interactions,…

Multiagent Systems · Computer Science 2026-03-03 Moritz Weckbecker , Jonas Müller , Ben Hagag , Michael Mulet

Most self-supervised learning (SSL) methods learn continuous visual representations by aligning different views of the same input, offering limited control over how information is structured across representation dimensions. In this work,…

Computer Vision and Pattern Recognition · Computer Science 2026-02-11 Kawtar Zaher , Ilyass Moummad , Olivier Buisson , Alexis Joly

Aligning large language models (LLMs) to preference data typically assumes a known link function between observed preferences and latent rewards (e.g., a logistic Bradley-Terry link). Misspecification of this link can bias inferred rewards…

Machine Learning · Computer Science 2026-02-03 Nathan Kallus

Large language models (LLMs) have shown remarkable adaptability to diverse tasks, by leveraging context prompts containing instructions, or minimal input-output examples. However, recent work revealed they also exhibit label bias -- an…

Computation and Language · Computer Science 2024-05-07 Yuval Reif , Roy Schwartz

Large language models (LLMs) are increasingly deployed in decision-support systems for high-stakes domains such as hiring and university admissions, where choices often involve selecting among competing alternatives. While prior work has…

Artificial Intelligence · Computer Science 2026-04-15 Haonan Yin , Shai Vardi , Vidyanand Choudhary

Scalable oversight studies methods of training and evaluating AI systems in domains where human judgment is unreliable or expensive, such as scientific research and software engineering in complex codebases. Most work in this area has…

Machine Learning · Computer Science 2024-10-22 Alex Mallen , Nora Belrose

Alignment techniques for LLMs rely on optimizing preference-based objectives -- where these preferences are typically elicited as ordinal, binary choices between responses. Recent work has focused on improving label quality or mitigating…

Artificial Intelligence · Computer Science 2025-08-13 Parker Whitfill , Stewy Slocum

A growing specter in the rise of machine learning is whether the decisions made by machine learning models are fair. While research is already underway to formalize a machine-learning concept of fairness and to design frameworks for…

Machine Learning · Computer Science 2020-09-28 Tao Zhang , Tianqing Zhu , Jing Li , Mengde Han , Wanlei Zhou , Philip S. Yu

The adoption of large language models (LLMs) is transforming the peer review process, from assisting reviewers in writing detailed evaluations to generating entire reviews automatically. While these capabilities offer new opportunities,…

Computers and Society · Computer Science 2026-04-29 Sai Suresh Macharla Vasu , Ivaxi Sheth , Hui-Po Wang , Ruta Binkyte , Mario Fritz

Integrated interpretability without sacrificing the prediction accuracy of decision making algorithms has the potential of greatly improving their value to the user. Instead of assigning a label to an image directly, we propose to learn…

Machine Learning · Computer Science 2021-04-13 Stephan Alaniz , Diego Marcos , Bernt Schiele , Zeynep Akata
‹ Prev 1 2 3 10 Next ›