Prompts & Provenance:
Re-situating Place in
Text-to-Image Models