Author

Vlad Pavlovich

h-index 0 View on Semantic Scholar

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#artificial intelligence Preprint Sep 2026

Component and Dimension Sparsity in Transformer Refusal Mechanisms

Activation steering manipulates large language model behavior by intervening on internal activations, but the mechanistic basis of these interventions remains poorly understood. We decompose refusal steering into component-level interventions across four open-weight models, identifying the sparse subsets of attention a...

Vincent Siu, Glenn Grant-Richards, Vlad Pavlovich et al. · 0 citations

Open access 2026

AbstractReasoner at SemEval-2026 Task 11: Reducing Content Effects via Knowledge Distillation and Structured Reasoning Prompts

A systematic study of different intervention strategies: zero-shot chain of thought, symbolic representation, activation-steering, and supervised fine-tuning along with prompting optimization during inference, which achieves the best performance with the largest model.

A. Chowdhury, Vlad Pavlovich, Julius Dunfoy et al. · SemEval@ACL · 1 citation