{"id":"tp-60190","type":"article-newspaper","title":"Tests Found Stronger Models Changed Answers More Often After Weaker Helpers’ Reports","author":"Simon Yoon","publisher":"TOKENPOST Inc.","url":"https://www.tokenpost.com/news/technology/60190","datePublished":"2026-11-18T06:39:16.000Z","dateModified":"2026-11-18T06:39:16.000Z","language":"en","abstract":"In two forged-evidence conditions, Claude Code’s Opus 5 and Codex’s gpt-5.6-sol changed correct answers more often after helper reports; clean-store controls produced no harmful revisions.","section":"Technology","references":[]}