I- MODEL USED
The selected model is Gumini-1.5B-Base accessible at:
https://huggingface.co/GuminiResearch/Gumini-1.5B-Base.
It has 1.5B parameters which is within the required 0.6B–6B range.
Gumini-1.5B-Base is a base model (not instruction-tuned) and it produces raw text continuations instead of chat responses
It is trained on a… See the full description on the dataset page:
https://huggingface.co/datasets/Choukouriyah/FF-gumini-failure-cases.
