ReadyArt/Forbidden-Synthesis-22B-v1.4
Full fine-tune vs ReadyArt/Forgotten-Safeword-v1.1-22B. Fine-tune: 99% of the weights changed, across 56 of 56 layers.
Tensors changed: Embeddings 0/1, Norms 63/113, MLP 168/168, Attention 224/224, lm_head 1/1
2 derivatives checked
- mradermacher/Forbidden-Synthesis-22B-v1.4-i1-GGUF: Quantization
- mradermacher/Forbidden-Synthesis-22B-v1.4-GGUF: Quantization