huggyllama/llama-30b + kaiokendev/SuperCOT-LoRA

Supercot was trained to work with langchain prompting.