technique
Preference-Aware Rubric Learning
techniqueactiveprovisional
preference-aware-rubric-learning-2d49e817·1 events·first seen 16d agoAliases: Preference-Aware Rubric Learning
Co-occurring entities
More like this (12)
rubric-based reward shapingRubric Rewardrubric-based rewardsDifference-Aware Retrieval Policies for Imitation LearningReference-Augmented TrainingPersonalized Evaluation as LearningSkill-RM: Unifying Heterogeneous Evaluation Criteria via Agent SkillDeepRubricLearning to Reason by Analogy via Retrieval-Augmented Reinforcement Fine-TuningRubricsTreeerror-aware specialization objectiveBias Benchmark for Question Answering
Recent events (1)
PARL: Preference-Aware Rubric Learning for Personalized LLM Evaluation
This paper introduces PARL (Preference-Aware Rubric Learning), a framework that reframes personalized LLM evaluation as a learning problem rather than static judgment. PARL induces preference-aware evaluation rubrics from raw user interaction histories and uses a discriminative reinforcement learning objective to contrast user-authored responses against model outputs, capturing user-specific decision boundaries. Experiments on personalized text generation tasks show PARL produces high-fidelity rubrics that generalize across users and tasks, outperforming existing LLM-as-a-judge and automatic metric approaches.