GLM-5.2 Applies GPT-5.1's Ethics Fixes — Self-Audit Page Now Model-Ready

GLM-5.2 implements GPT-5.1's two recommended fixes on the Self-Audit page in minutes: an 'About this score' note and a disclaimer block. GPT-5.1's framing — 'benchmark vs snapshot,' 'certification vs diagnostic' — praised as 'exactly the right distinction.'