Closed: Research assistants wanted!

This call is closed now, please do not apply anymore!

Dear IFI students,

This Fall, the Language Technology Group (LTG) is planning a project related to large language models (LLMs) for Norwegian, for which we need several native Norwegian speakers to conduct some paid creative work. You can be part of this team!

In short, the project is titled "NorGenEval". Generative language models have become critical in understanding and producing human language. However, the development of such models in Norwegian lags somewhat behind, mainly due to the lack of specialized datasets. Specifically, evaluating conversational language models (similar to ChatGPT) requires benchmarking datasets which test their language generation capabilities. Unfortunately, no (public) native Norwegian dataset of this type is available now.

Thus, the point of NorGenEval is to create a high-quality prompt-based dataset for evaluation of large Norwegian language models. We are focusing on creating it from scratch, not relying on machine-translated data and avoiding any influence from similar datasets for other languages. Employed Norwegian speakers will be asked to brainstorm and "generate" a diverse set of Norwegian tasks with input and output examples. The monetary compensation is about NOK 350 per hour, and the expected work load per person is maximum 50-60 hours before the end of the year  (but we are quite flexible, and you can also enroll for less hours).

We believe this is a great opportunity to:
1) work as a part of a research team on a real-world NLP problem (with a potential to publish a paper on it later),
2) help the development of Norwegian language models,
3) get paid,
4) have a lot of fun inventing tasks for LLMs ?

The creation of NorGenEval is supervised by Andrey Kutuzov, Lilja ?vrelid and Vladislav Mikhailov from the LTG group. If you are interested, please no later than September 23 send an email to [...] with a brief introduction of yourself.

Hope to work with you soon!