Elisabetta Rocchetti

Active

Research focus

I am focusing on interpretability, exploring different approaches for modelling neural networks, and understanding how models organise concepts in their latent space.

Open to collaboration

Yes

Looking for

volunteers and independent contributors, or collaborators from other research labs

Contact

www.linkedin.com/in/elisabetta-rocchetti-70a953143

Join the Damaqu community to see this team's full profile.

Papers on Damaqu (11)

Processing and Representation of Linguistic Properties in Large Language Models
other
Augustine of Hippo at SemEval-2023 Task 4: An Explainable Knowledge Extraction Method to Identify Human Values in Arguments with SuperASKE
other
A Discussion on Open Issues Regarding Human Value Detection in Arguments
other
Causal Mediation Analysis for Interpreting Large Language Models
other
The Cow of Rembrandt Analyzing Artistic Prompt Interpretation in Text-to-Image Models
other
Refusal Beyond a Single Direction: A Preliminary Comparison of Diff-in-Means and INLP
arxiv· 11 Jun 2026
Unveiling Transformer Perception by Exploring Input Manifolds
other· 23 Apr 2026
How LLMs Follow Instructions: Skillful Coordination, Not a Universal Mechanism
arxiv· 7 Apr 2026
Modeling Transformers as complex networks to analyze learning dynamics
arxiv· 18 Sept 2025
How Instruction-Tuning Imparts Length Control: A Cross-Lingual Mechanistic Analysis
arxiv· 2 Sept 2025
What's Taboo for You? - An Empirical Evaluation of LLMs Behavior Toward Sensitive Content
arxiv· 31 Jul 2025