BM-TOK
Balancing multilingual token vocabularies
How to build the vocabulary of a multilingual language model so that less-resourced languages and non-Latin scripts are represented fairly.
David worked on interpretability of language models and on using them for information retrieval in specialised documents. He holds a PhD in NLP from Sorbonne Nouvelle and Université Paris Cité, on how pre-trained language models represent negation.
Projects
LPHI-mRAG
Language preference and hallucination in multilingual retrieval-augmented generation
Do language models prefer some languages over others when answering from documents? We test this for chatbots that answer questions from a collection of documents in several languages.
UBS–IDSIA AI Lab
UBS–IDSIA AI Lab collaboration
The NLP group participates in the collaboration across different NLP-related projects.
Publications with the group
- Are Language Models Script-Aware?
- Script Choice in LLMs: Evidence for Late-Layer Commitment
- Are the LLMs Capable of Maintaining at Least the Language Genus?
- Presenting LLMs' Collective Intelligence Approach for Multilingual Hallucination Detection