{"type":"link","version":"1.0","title":"Per-step accuracy metrics mask trajectory-level failure in stateful multi-step agent tasks","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-per-step-accuracy-masks-trajectory-failure","thumbnail_url":"https://ai-arch.pages.dev/og/risk-per-step-accuracy-masks-trajectory-failure.png","thumbnail_width":1200,"thumbnail_height":630}