GEITje 7B Ultra: A Conversational Model for Dutch
This provides a more capable Dutch conversational model for Dutch speakers, though it's incremental as it builds on existing adaptation work.
The researchers tackled the problem of limited Dutch conversational AI by extending the GEITje model (derived from English-based Mistral 7B) through supervised finetuning on new synthetic conversational datasets and preference alignment on synthetic feedback, making both models and datasets openly available.
Language models have rapidly evolved, predominantly focusing on English while often neglecting extensive pretraining in other languages. This approach has required initiatives to adapt powerful, English-centric models to other linguistic contexts through finetuning. For Dutch, such a recent endeavour is ``GEITje'' a model originally derived from the English-based Mistral 7B. Building on this fundamental work, the current research extends the capabilities of GEITje by supervised finetuning on newly created high-quality synthetic conversational datasets, along with an additional preference alignment procedure on a synthetic feedback dataset. Both the developed models and the created datasets are openly available.