46 6 days ago

DASD-4B-Thinking is a compact yet capable 4B dense language model specialized in long chain-of-thought (Long-CoT) reasoning across mathematics, code generation, and scientific reasoning. This version has overfitting! Avoid it.

4b
39a81f781272 · 43B
{
"num_ctx": 4096,
"temperature": 1,
"top_p": 1
}