The role involves preparing and validating historical corpora, fine-tuning open-weight language models, building attribution pipelines, and running systematic model tests to support a research benchmark release. Candidates must hold a doctorate in a computational field, possess experience with transformer language models in Python, and have knowledge of interpretability methods such as SHAP or probing. This position covers a four-year project duration.
(KI generierte Zusammenfassung)
Inserat ansehen