Biography
Research scientist at RIKEN and Tohoku University. Aiming to understand if/how language models understand language. Specific research interests are at the interesection of knowledge bases and language models, including questions such as how world knowledge is represented in language models, how internal representations relate to model behavior, and if language models can “know” anything at all. Prior to joining the language understanding team at RIKEN, Benjamin received a PhD in computational linguistics from Heidelberg University.
Interests
- Knowledge representation in language models
- Interpretability, internal representations of language models
- Information extraction
- Tokenization, multilingual subword methods
Education
-
PhD in Computational Linguistics, 2018
Heidelberg University
-
Magister in Computational Linguistics, 2015
Heidelberg University
