Journal Article
Scaling Laws for Data Filtering-Data Curation <i>cannot</i> be Compute Agnostic
T-MARS: IMPROVING VISUAL REPRESENTATIONS BY CIRCUMVENTING TEXT FEATURE LEARNING
T-MARS: Improving Visual Representations by Circumventing Text Feature Learning.
Test-Time Adaptation Induces Stronger Accuracy and Agreement-on-the-Line
Text Descriptions are Compressive and Invariant Representations for Visual Learning
THE UPDATE-EQUIVALENCE FRAMEWORK FOR DECISION-TIME PLANNING
The Update-Equivalence Framework for Decision-Time Planning.
TOFU: A Task of Fictitious Unlearning for LLMs.
Transformers to SSMs: Distilling Quadratic Knowledge to Subquadratic Models
UNDERSTANDING AUGMENTATION-BASED SELF-SUPERVISED REPRESENTATION LEARNING VIA RKHS APPROXIMATION AND REGRESSION
Understanding Hallucinations in Diffusion Models through Mode Interpolation
UNDERSTANDING PROMPT ENGINEERING MAY NOT REQUIRE RETHINKING GENERALIZATION
Understanding prompt engineering may not require rethinking generalization.
WHY IS SAM ROBUST TO LABEL NOISE?
(CERTIFIED!!) ADVERSARIAL ROBUSTNESS FOR FREE!
A UNIFIED APPROACH TO REINFORCEMENT LEARNING, QUANTAL RESPONSE EQUILIBRIA, AND TWO-PLAYER ZERO-SUM GAMES