{"type":"link","version":"1.0","title":"Critic agents trained on corrected debate trajectories carry most of the gain on hard reasoning, little on easy","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/trained-critic-agents-carry-the-gain-on-hard-reasoning","thumbnail_url":"https://ai-arch.pages.dev/og/trained-critic-agents-carry-the-gain-on-hard-reasoning.png","thumbnail_width":1200,"thumbnail_height":630}