Daemontatox/Llama3.3-70B-CogniLink
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
5 derivatives checked
- mradermacher/Llama3.3-70B-CogniLink-i1-GGUF: Quantization
- mradermacher/Llama3.3-70B-CogniLink-GGUF: Quantization
- tensorblock/Llama3.3-70B-CogniLink-GGUF: Quantization
- TareksGraveyard/Perspicacious-LLaMa-70B: Full fine-tune
- schonsense/70B_thinkthonk: Full fine-tune