ParentFull threadcowthulhu·Does anyone know how well a fine-tuned Llama model will do compared to GPT 4 on complex tasks?View on HN