Publications

Simulating the Emergence of Differential Case Marking with Communicating Neural-Network Agents
Cite DOI URL
Reading Between the Prompts: How Stereotypes Shape LLM's Implicit Personalization
Cite DOI URL
Pointwise Mutual Information as a Performance Gauge for Retrieval-Augmented Generation
Recent work suggests that large language models enhanced with retrieval-augmented generation are easily influenced by the order in …
Cite URL
On the Consistency of Multilingual Context Utilization in Retrieval-Augmented Generation
Retrieval-augmented generation (RAG) with large language models (LLMs) has demonstrated strong performance in multilingual …
Cite URL
Steering Large Language Models for Machine Translation Personalization
Cite URL
QE4PE: Word-level Quality Estimation for Human Post-Editing
Cite arXiv URL
MultiBLiMP 1.0: A Massively Multilingual Benchmark of Linguistic Minimal Pairs
Cite arXiv URL
Child-Directed Language Does Not Consistently Boost Syntax Learning in Language Models
Cite arXiv URL
Non Verbis, Sed Rebus: Large Language Models Are Weak Solvers of Italian Rebuses
Cite URL
EurekaRebus - Verbalized Rebus Solving with LLMs: A CALAMITA Challenge
Cite URL
Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation
Cite DOI URL
The SIFo Benchmark: Investigating the Sequential Instruction Following Ability of Large Language Models
Following multiple instructions is a crucial ability for large language models (LLMs). Evaluating this ability comes with significant …
Cite DOI URL
NeLLCom-X: A Comprehensive Neural-Agent Framework to Simulate Language Learning and Group Communication
Cite URL
Multi-property Steering of Large Language Models with Dynamic Activation Composition
Cite DOI URL
Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation
Ensuring the verifiability of model answers is a fundamental challenge for retrieval-augmented generation (RAG) in the question …
Cite DOI URL
BabyLM Challenge: Exploring the effect of variation sets on language model training efficiency
Cite URL
Democratizing Advanced Attribution Analyses of Generative Language Models with the Inseq Toolkit
PDF Cite
Encoding of lexical tone in self-supervised models of spoken language
Cite DOI URL
DecoderLens: Layerwise Interpretation of Encoder-Decoder Transformers
In recent years, several interpretability methods have been proposed to interpret the inner workings of Transformer models at different …
Cite URL
Quantifying the Plausibility of Context Reliance in Neural Machine Translation
Cite URL
MBBQ: A Dataset for Cross-Lingual Comparison of Stereotypes in Generative LLMs
Cite URL
IT5: Text-to-text Pretraining for Italian Language Understanding and Generation
We introduce IT5, the first family of encoder-decoder transformer models pretrained specifically on Italian. We document and perform a …
Cite URL
Are Character-level Translations Worth the Wait? Comparing ByT5 and mT5 for Machine Translation
Cite DOI URL
Neural-agent Language Learning and Communication: Emergence of Dependency Length Minimization
Cite
Endowing neural language learners with human-like biases: A case study on dependency length minimization
Cite
Automatic Text Simplification: A Comparative Study in Italian for Children with Language Disorders
Cite
A Functional Trade-off between Prosodic and Semantic Cues in Conveying Sarcasm
Cite
Cross-Lingual Consistency of Factual Knowledge in Multilingual Language Models
Multilingual large-scale Pretrained Language Models (PLMs) have been shown to store considerable amounts of factual knowledge, but …
Cite DOI URL
Communication Drives the Emergence of Language Universals in Neural Agents: Evidence from the Word-order/Case-marking Trade-off
Cite DOI URL
Wave to Syntax: Probing spoken language models for syntax
Cite DOI URL
RAMP: Retrieval and Attribute-Marking Enhanced Prompting for Attribute-Controlled Translation
Cite URL
Inseq: An Interpretability Toolkit for Sequence Generation Models
Cite URL
Contrastive Language-Image Pre-training for the Italian Language
PDF Cite
DivEMT: Neural Machine Translation Post-Editing Effort Across Typologically Diverse Languages
We introduce DivEMT, the first publicly available post-editing study of Neural Machine Translation (NMT) over a typologically diverse …
Cite DOI URL
Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
Cite DOI URL
A Primer on the Inner Workings of Transformer-based Language Models
Cite URL