SicariusSicariiStuff/Negative_LLAMA_70B
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
- bartowski/Negative_LLAMA_70B-GGUF: Quantization
- mradermacher/Negative_LLAMA_70B-i1-GGUF: Quantization
- mradermacher/Negative_LLAMA_70B-GGUF: Quantization
- knifeayumu/Negative-Anubis-70B-v1: LoRA-style
- ReadyArt/Forgotten-Safeword-70B-v5.0: Full fine-tune
- altomek/RE-70B-AS3D: Full fine-tune
- NexesMess/Llama_3.3_70b_L3.3-NegativeLlama_abliterated_fusion: LoRA-style
- sh0ck0r/Forgotten-Safeword-70B-v5.0-heretic: Full fine-tune
- Nexesenex/Llama_3.3_70b_L3.3-NegativeLlama_abliterated_fusion_norm: LoRA-style
1 other repos hold byte-for-byte identical weight files.