Tag
This paper analyzes how localizing edits in human corrections can amplify relative selection bias in model training, providing theoretical insights and empirical evidence from machine translation experiments.
This paper proposes Transfer-Aware Curriculum (TAC), a bandit-style online curriculum for multi-domain RLVR that prioritizes domains whose updates benefit other domains using gradient-geometry alignment. TAC improves macro-averaged accuracy on Qwen3-1.7B and Llama3.2-3B over fixed and learnability-only curricula.