University of Cambridge > Talks.cam > NLIP Seminar Series > Assessing language-specific capabilities of LLMs: Lessons from Swedish NLP

  Browse A-Z

Log in

University Account External (via Google)

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹī¸ Information

Assessing language-specific capabilities of LLMs: Lessons from Swedish NLP

Download to your calendar using vCal

If you have a question about this talk, please contact Suchir Salhan .

Abstract: In this talk, I discuss benchmarking and interpreting large language models in the context of Swedish. I present a selection of work from my PhD thesis, which analyze LLMs Swedish-specific capabilities in different areas: English-Swedish language transfer, multi-task benchmarking on Swedish NLU and targeted analysis of a specific case of Swedish linguistic variation. I outline some of the challenges that arise when trying to assess the language-specific capabilities of LLMs, some lessons I’ve learned throughout from my work and give a future outlook.

Bio: Felix Morger is a software engineer and computational linguist based in Gothenburg, Sweden. He recently defended his PhD thesis, titled “In the minds of stochastic parrots: Benchmarking, evaluating and interpreting large language models”. His main research interests are in benchmarking and model interpretability of large language models with a large focus on Swedish.

This talk is part of the NLIP Seminar Series series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Average on weekdays: 9012.53109656 kWh JDB Seminar Room,, Engineering Design Centre, CUED Hybrid: LT6, Engineering Department or Zoom: https://cam-ac-uk.zoom.us/j/83363886282?pwd=eCJD8vmmxI29vSZMX92lmQjWWUgZgw.1

Other talks

Title to be confirmed TOWARD AFFORDABLE PREVENTIVE HEALTHCARE: How basic science addresses grand challenges Title TBC Title TBC Title to be confirmed Group Presentation 1
 

Š 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity