reedmayhew

/

Llama-3.1-8B-claude-3.7-sonnet-reasoning-distilled

like 4

Hello!

This model does this for some reason under certain circumstances. It has to do with the content inputted.

I still have not figured out why. If possible, I'd highly suggest switching to:
reedmayhew/gemma3-12B-claude-3.7-sonnet-reasoning-distilled

The gemma3 models are very stable and do not exhibit these kinds of errors and confirm to the fine-tuning better, which is why I've stopped fine-tuning the llama models.

Thanks!

Thanks, ill test it tomorrow and give feedback.