SOURCE-LINKED INTELLIGENCE

From Preferences to Principles: Rubric-Based Alignment for Grounded Knowledge Answers

arXiv · AI, language, vision and robotics · article · Aug 24, 2026 · UTC

Designing effective reward signals for open-domain question answering is challenging because high-quality responses must simultaneously satisfy multiple aspects of answer quality that are difficult to capture with a holistic scalar objective. We introduce a rubric-based reward framework that generates query-specific rubrics grounded in retrieved evidence and decomposed into multiple quality dimensions, providing fine-grained supervision during post-training. Averaged across three evaluation axes (composition, grounding, and instruction-following), our approach improves over the instruction-tun

Read original source ↗ Open in workspace

recordType

paper

region

Global