Associated with the Linz Institute of Technology and the Institute of Computational Perception @ JKU Linz, Austria

On the Applicability of Dimensionality Reduction Techniques as Aggregation Methods to Enhance Feature Attribution Faithfulness of LLMs

Abstract


Citation

Nico Filzmoser
On the Applicability of Dimensionality Reduction Techniques as Aggregation Methods to Enhance Feature Attribution Faithfulness of LLMs
Advisor(s): Markus Schedl,
Johannes Kepler University Linz, Master's Thesis, 2025.

BibTeX

@misc{NicoFilzmoser2025master-thesis,
    title = {On the Applicability of Dimensionality Reduction Techniques as Aggregation Methods to Enhance Feature Attribution Faithfulness of LLMs},
    author = {Nico Filzmoser},
    school = {Johannes Kepler University Linz},
    year = {2025}
}

Resources