Comment Comment by amelius Does this train on LLM output, or is this more like iterative self prompt improvement? parent amelius 5 months ago
