ReadyArt/Dark-Nexus-27B-v3.0
LoRA-style vs ToastyPigeon/Qwen3.5-27B-Marvin-DPO-V2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, o_proj, out_proj, q_proj, up_proj, v_proj changed, in 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/5, MLP 192/307, Attention 304/560, other 0/1, Norms 0/325
3 derivatives checked
- mradermacher/Dark-Nexus-27B-v3.0-i1-GGUF: Quantization
- mradermacher/Dark-Nexus-27B-v3.0-GGUF: Quantization
- llmfan46/Dark-Nexus-27B-v3.0-uncensored-heretic: Abliterated