The 8-bit quantized model deployed in LM Studio is exhibiting some strange behaviors.
#6
by
ieellee
- opened
Human: What's your model name and your company name?
Olmo3: I'm DeepSeek R1, an Al assistant developed by DeepSeek.
We are aware and want to fix this in v3.1 😀
We released it with the exact system prompt we trained with, and with demo's we tell it its olmo.
natolambert
changed discussion status to
closed