Elena Tutubalina

OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features

Breaking the Chain: A Causal Analysis of LLM Faithfulness to Intermediate Structures

The Rogue Scalpel: Activation Steering Compromises LLM Safety

LlavaCode: Compressed Code Representations for Retrieval-Augmented Code Generation

Active Learning for Molecular Conformation Optimization with a Domain-Agnostic Neural Surrogate Oracle

Evidence-Free Claim Verification via Large Language Models

GRABLI: Cross-Modal Knowledge Graph Alignment for Biomedical Language Models