{"type":"link","version":"1.0","title":"A router trained on its own logged choices imitates its past mistakes and drifts, unless the data is labeled with environment outcomes","author_name":"AI Archs","author_url":"https://ai-arch.pages.dev","provider_name":"AI Archs","provider_url":"https://ai-arch.pages.dev","url":"https://ai-arch.pages.dev/n/risk-router-self-poisoning-from-imitating-logged-decisions","thumbnail_url":"https://ai-arch.pages.dev/og/risk-router-self-poisoning-from-imitating-logged-decisions.png","thumbnail_width":1200,"thumbnail_height":630}