3

Steering Vectors are an Adversarial Attack Surface
Activation steering has become a popular way to control Large Language Model (LLM) behavior without fine-tuning. Since the technique is …
Abzal Aidakhmetov, Donato Crisostomi, Tommaso Mencattini, Adrian Robert Minut, Iacopo Masi, Emanuele Rodolà
arXiv
Steering Vectors are an Adversarial Attack Surface
PHALAR: Phasors for Learned Musical Audio Representations
Stem retrieval, the task of matching missing stems to a given audio submix, is a key challenge currently limited by models that discard …
Davide Marincione, Michele Mancusi, Giorgio Strano, Luca Cerovaz, Donato Crisostomi, Roberto Ribuoli, Emanuele Rodolà
arXiv
PHALAR: Phasors for Learned Musical Audio Representations
Post-AGI Science and Society Workshop
The Post-AGI Science and Society Workshop at ICLR 2026 asks what comes next if AGI becomes ubiquitous, reliable, and affordable, and …
Donato Crisostomi, Andrea Santilli, Pratyusha Sharma, Valentina Pyatkin, Zorah Lähner, Emanuele Rodolà
URL
Post-AGI Science and Society Workshop
Multi-objective Evolutionary Merging Enables Efficient Reasoning Models
Reasoning models have demonstrated remarkable capabilities in solving complex problems by leveraging long chains of thought. However, …
Mario Iacobelli, Adrian Robert Minut, Tommaso Mencattini, Donato Crisostomi, Andrea Santilli, Iacopo Masi, Emanuele Rodolà
arXiv
Multi-objective Evolutionary Merging Enables Efficient Reasoning Models
Multi-Way Representation Alignment
The Platonic Representation Hypothesis suggests that independently trained neural networks converge to increasingly similar latent …
Akshit Achara, Tatiana Gaintseva, Mateo Mahaut, Pritish Chakraborty, Viktor Stenby Johansson, Melih Barsbey, Emanuele Rodolà, Donato Crisostomi
arXiv
Multi-Way Representation Alignment
Membership and Dataset Inference Attacks on Large Audio Generative Models
Generative audio models, based on diffusion and autoregressive architectures, have advanced rapidly in both quality and expressiveness. …
Jakub Proboszcz, Paweł Kochański, Karol Korszun, Katarzyna Stankiewicz, Donato Crisostomi, Giorgio Strano, Emanuele Rodolà, Kamil Deja, Jan Dubiński
arXiv
Membership and Dataset Inference Attacks on Large Audio Generative Models
Model Merging Improves Zero-Shot Generalization in Bioacoustic Foundation Models
Foundation models capable of generalizing across species and tasks represent a promising new frontier in bioacoustics, with NatureLM …
Davide Marincione, Donato Crisostomi, Roberto Dessì, Emanuele Rodolà, Emanuele Rossi
arXiv GitHub
Model Merging Improves Zero-Shot Generalization in Bioacoustic Foundation Models
Language Models are Injective and Hence Invertible
Transformer components such as non-linear activations and normalization are inherently non-injective, suggesting that different inputs …
Giorgios Nikolaou, Tommaso Mencattini, Donato Crisostomi, Santilli, Yannis Panagakis, Emanuele Rodolà
Cite arXiv Thread
Language Models are Injective and Hence Invertible
Two-Scale Latent Dynamics for Recurrent-Depth Transformers
Recurrent-depth transformers scale test-time compute by iterating latent computations before emitting tokens. We study the geometry of …
Francesco Pappone, Donato Crisostomi, Emanuele Rodolà
arXiv
Two-Scale Latent Dynamics for Recurrent-Depth Transformers
On Task Vectors and Gradients
Task arithmetic has emerged as a simple yet powerful technique for model merging, enabling the combination of multiple finetuned models …
Luca Zhou, Daniele Solombrino, Donato Crisostomi, Maria Sofia Bucarelli, Giuseppe Alessio D'Inverno, Fabrizio Silvestri, Emanuele Rodolà
arXiv
On Task Vectors and Gradients