Papers by Gabriele Sarti
Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent research has shown that self-citing large language models (LLMs) fail to faithfully reflect their context usage throughout the generation process. |
| Approach: | They propose a plug-and-play approach using model internals for faithful answer attribution in RAG applications that detects context-sensitive answer tokens and pairs them with retrieved documents contributing to their prediction. |
| Outcome: | The proposed approach achieves citation quality and efficiency comparable to self-citation while allowing for a finer-grained control of attribution parameters. |
DivEMT: Neural Machine Translation Post-Editing Effort Across Typologically Diverse Languages (2022.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in neural language modeling and multilingual training have prompted widespread adoption of machine translation (MT) technologies across an unprecedented range of world languages. |
| Approach: | They propose to use a dataset to assess the impact of two state-of-the-art NMT systems, Google Translate and the multilingual mBART-50 model, on translation productivity. |
| Outcome: | The proposed model is faster than translation from scratch, but the magnitude of productivity gains varies widely across systems and languages. |
Steering Large Language Models for Machine Translation Personalization (2026.eacl-long)
Copied to clipboard
| Challenge: | Recent advances in interpretability research have highlighted the effectiveness of steering methods for MT personalization. |
| Approach: | They examine steering strategies for personalizing automatic translations when few examples are available. |
| Outcome: | The proposed steering methods yield higher inference-time computational efficiency than prompting approaches. |
IT5: Text-to-text Pretraining for Italian Language Understanding and Generation (2024.lrec-main)
Copied to clipboard
| Challenge: | Xue et al., 2022) use the text-to-text paradigm to train multilingual models. |
| Approach: | They introduce the first family of encoder-decoder transformer models pretrain specifically on Italian and introduce the ItaGen benchmark to evaluate the models' performance. |
| Outcome: | The proposed model outperforms models with multilingual baselines and the original model on English data. |
Interpreto: An Explainability Library for Transformers (2026.acl-demo)
Copied to clipboard
Antonin Poché, Thomas Mullor, Gabriele Sarti, Frédéric Boisnard, Corentin Friedrich, Charlotte Claye, Francois Hoofd, Raphael Bernas, Nicholas Asher, Celine Hudelot, Fanny Jourdan
| Challenge: | Interpreto is an open-source Python library for interpreting HuggingFace language models . it provides attribution methods and concept-based explanations . documentation or metrics are sometimes missing due to the complexity of the pipeline . |
| Approach: | Interpreto is an open-source Python library for interpreting HuggingFace language models . it provides attribution methods and concept-based explanations . authors welcome issues and pull requests . |
| Outcome: | Interpreto is an open-source Python library for interpreting HuggingFace language models . it provides attribution methods and concept-based explanations . the library welcomes issues and pull requests . |
Inseq: An Interpretability Toolkit for Sequence Generation Models (2023.acl-demo)
Copied to clipboard
| Challenge: | Recent studies focused on classification tasks while largely overlooking generation settings due to a lack of dedicated tools. |
| Approach: | They propose to use Inseq to democratize access to interpretability analyses of sequence generation models by enabling intuitive extraction of models’ internal information and feature importance scores for popular decoder-only and encoder-decoder Transformers architectures. |
| Outcome: | The proposed library can extract models’ internal information and feature importance scores for popular decoder-only and encoder-decoder Transformers architectures. |
RAMP: Retrieval and Attribute-Marking Enhanced Prompting for Attribute-Controlled Translation (2023.acl-short)
Copied to clipboard
| Challenge: | Attribute-controlled translation (ACT) is a subtask of machine translation that involves controlling stylistic or linguistic attributes (like formality and gender) of outputs. |
| Approach: | They propose a new approach to attribute-controlled translation that leverages multilingual language models to perform ACT in few-shot and zero-shot settings. |
| Outcome: | The proposed approach improves generation accuracy over the standard prompting approach in both zero-shot and few-shot settings. |
Unsupervised Word-level Quality Estimation for Machine Translation Through the Lens of Annotators (Dis)agreement (2025.emnlp-main)
Copied to clipboard
| Challenge: | Modern WQE techniques rely on expensive inference with large language models or ad-hoc training with large amounts of human-labeled data. |
| Approach: | They propose to use word-level quality estimation to identify translation errors from the inner workings of translation models to quantify the impact of human label variation on metric performance. |
| Outcome: | The proposed methods identify translation errors from the inner workings of translation models using human labels. |
DecoderLens: Layerwise Interpretation of Encoder-Decoder Transformers (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing interpretability methods have been proposed to interpret the inner workings of Transformer models at different levels of precision and complexity. |
| Approach: | They propose a method to analyze encoder-decoder Transformers by using the decoder module Model Output encoder to cross-attend representations of intermediate encoder activations instead of using the default output. |
| Outcome: | The proposed method maps uninterpretable representations to human-interpreted sequences of words or symbols, shedding new light on the information flow in this popular but understudied class of models. |