Introducing LLM Optimize, a toy proof-of-concept library for LLM-guided blackbox optimization using GPT-4. Perform optimization on problems beyond the usual numerical methods, such as code-based AutoML and natural language rubric-based optimization.
Check it out: https://github.com/sshh12/llm_optimize