CrucibleLab/L3.3-70B-Loki-V2.0
LoRA-style vs meta-llama/Llama-3.3-70B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 80 of 80 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/161, MLP 240/240, Attention 320/320, lm_head 0/1
9 derivatives checked
- mradermacher/L3.3-70B-Loki-V2.0-i1-GGUF: Quantization
- mradermacher/L3.3-70B-Loki-V2.0-GGUF: Quantization
- Silicone-Moss/CrucibleLab-L3.3-70B-Loki-V2.0-Heretic-Uncensored-GGUF: Abliterated
- CrucibleLab/L3.3-70B-Loki-V2.0-GGUF: Quantization
- Naphula/Cthulhu-70B-v1: Full fine-tune
- Darkknight535/Void-Citrus-L3.3-70B: Full fine-tune
- Silicone-Moss/CrucibleLab-L3.3-70B-Loki-V2.0-Heretic-Uncensored: Abliterated
- hackoffice/Loki-V2.0-Heretic-Uncensored: Abliterated
- ReadyArt/Omega-Evolution-70B-v2.1: LoRA-style