When your LLM prompts stop working as well — the model has been updated, training data has shifted, or the context window is now choked with irrelevant history.
"My eval prompts broke overnight. Classic prompt rot."
No comments yet — say something.
Add your own interpretation of "prompt rot".