https://comp30027-recipe-time.vercel.app https://comp30027-recipe-time.vercel.app/predict https://comp30027-recipe-time.vercel.app/models https://comp30027-recipe-time.vercel.app/evaluation https://comp30027-recipe-time.vercel.app/llm-eval https://comp30027-recipe-time.vercel.app/stacking https://comp30027-recipe-time.vercel.app/data https://comp30027-recipe-time.vercel.app/methods https://comp30027-recipe-time.vercel.app/ai-log https://comp30027-recipe-time.vercel.app/about https://comp30027-recipe-time.vercel.app/tour https://comp30027-recipe-time.vercel.app/methods/decisions/DR-001-rebuild-from-the-report https://comp30027-recipe-time.vercel.app/methods/decisions/DR-002-label-thresholds https://comp30027-recipe-time.vercel.app/methods/decisions/DR-003-stacking https://comp30027-recipe-time.vercel.app/methods/decisions/DR-004-byok-llm-evaluation-harness https://comp30027-recipe-time.vercel.app/methods/decisions/DR-005-how-uncertainty-is-reported https://comp30027-recipe-time.vercel.app/methods/decisions/DR-006-llm-test-primary-comparison-and-truncation