Karen Ullrich

OpenApps: Simulating Environment Variations to Measure UI Agent Reliability

How reinforcement learning after next-token prediction facilitates learning

LLM Output Homogenization is Task Dependent

From Concepts to Components: Concept-Agnostic Attention Module Discovery in Transformers

Cross-Tokenizer Likelihood Scoring Algorithms for Language Model Distillation

Simple Optimizers for Convex Aligned Multi-Objective Optimization

A Single Character Can Make or Break your LLM Evals

Exact Byte-Level Probabilities from Tokenized Language Models for FIM-Tasks and Model Ensembles

Aligned Multi Objective Optimization