CoolFace
Modelpublic

Zzyy2000/qwen3-4b-midtrain-v2-reasoning-only

sourceHugging Faceotherupdated 2mo agoView on Hugging Face
0likes5downloads
Model Card

Qwen3-4B Safety Midtrain ? v2 Reasoning Only

Off-policy baseline midtraining checkpoint initialized from Qwen3-4B-Base using the v2 chat-reasoning corpus. It is not initialized from the non-Base Qwen3-4B instruct model.