{"type":"link","version":"1.0","title":"Self-judged data improves a model through best-vs-worst preference pairs; fine-tuning on its top-scored outputs alone gave no gain","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/self-judged-training-needs-contrastive-pairs-not-curated-positives","thumbnail_url":"https://ai-arch.pages.dev/og/self-judged-training-needs-contrastive-pairs-not-curated-positives.png","thumbnail_width":1200,"thumbnail_height":630}