hauschke@osma At #SWIB23 we discussed with @SemAntiKast about some tests you did with LLM based classification, if I remember correctly. Do I remember correctly? If yes, did you document your tests, and did you try different approaches (reasoning models etc.)?
Osma Suominen@hauschke @SemAntiKast Hmm, now I'm not sure what discussion you're referring to.
We've mainly tried LLMs for metadata extraction, not classification as such. I gave a talk about it at #SWIB24 with Pierre Beauguitte from Norway (and also a lightning talk at #SWIB23 ).
Code for the experiments can be found here: https://github.com/NatLibFi/FinGreyLit
Most experiments involve fine-tuning smallish LLMs. Haven't tried reasoning models yet.
