Xirui1208/readall-readtwice-14b-stage2-localstep15-globalstep30-20260926

Partial fine-tune vs Xirui1208/readall-readtwice-14b-rl-step15-20260925. Partial change: 89% of the weights changed (attention, mlp, norms), in 48 of 48 layers.

Tensors changed: Embeddings 0/1, MLP 144/144, Attention 335/336, Norms 85/97, lm_head 0/1

On the Hub · Downloads on Model Pulse