automerger/Experiment26Pastiche-7B
Full fine-tune vs CorticalStack/pastiche-crown-clown-7b-dare-dpo. Fine-tune: 97% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 65/65, MLP 93/96, Attention 124/128, lm_head 1/1
4 derivatives checked
- MaziyarPanahi/Experiment26Pastiche-7B-GGUF: Quantization
- MaziyarPanahi/Experiment26Yam_Experiment26Pastiche: Full fine-tune
- MaziyarPanahi/M7Yamshadowexperiment28_Experiment26Pastiche: Full fine-tune
- MaziyarPanahi/Experiment28M7_Experiment26Pastiche: Full fine-tune