drwlf/DrMedra4B-abliterated

Full fine-tune vs google/medgemma-4b-it. Fine-tune: 90% of the weights changed, across 34 of 34 layers.

Tensors changed: Embeddings 1/4, Norms 0/316, MLP 102/210, Attention 136/352, other 0/1

2 derivatives checked

• mradermacher/DrMedra4B-abliterated-GGUF: Same weights, converted (fewer layers)
• mradermacher/DrMedra4B-abliterated-i1-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse