ProCreations/deepicml-demystifying-llm-as-a-judge-analytical-anvg7n

Audit Remark 3 threshold coefficients and finite k

Add LLM-as-a-judge model simulator and raw results

Identify and validate the analytic model, then test Remarks 3, 4 and 6: finite k* under misspecification, t*(k) ~ (1-2/k), and the -2 vs -1 scaling exponents

Replace with exact three-claim LLM-judge scaling audit

Update metadata with huggingface_hub

Update logbook: Direct evidence mapping

initial commit