deepseek-ai

/

DeepSeek-Prover-V2-7B

like 115

Follow

DeepSeek 85.9k

In subsection 2.3 of the paper is mentioned the step Distillation step of 671B into 7B

Is the released model the result of this distillation step?

Thanks :)

Curious, if you tried learning smaller models with CoT data - e.g. 7B/32B/235B?