Fix in load_llm.py (#1508)

Fixed an issue where the "proxy" setting was passed to the PublicOpenAPI constructor instead of the  "api_base" parameter, disabling the use of on-premise OpenAI-based LLM servers

Co-authored-by: Alonso Guevara <[email protected]>
Co-authored-by: Josh Bradley <[email protected]>

Author: 3 people