Supplementary Materials: How Large Language Models Source Brand Reputation Knowledge - A Cross-Industry Attribution Analysis
收藏官方服务：
资源简介：
Version 2.0 corrects four errors in version 1.0. The analysis frame carried 72 duplicated response rows whose derived measures were copied from sibling responses; deduplicating gives 1,368 responses balanced at 171 per sector. The retrieval-augmented arm recorded no citation URLs; it was re-collected on 26 July 2026 with the payload captured, giving 1,146 citations across 415 domains. RepTrak dimension coding, claimed but never performed, is now included for 288 responses. The periodicity analysis is withdrawn: 720 of its 780 comparisons compared a run to itself. Version 1.0 remains public and permanent.
创建时间：
2026-07-26
