L. Ma | Dezhi Yuhttps://halfrost.me/authors/l.-ma/L. MaHugoBlox Kit (https://hugoblox.com)en-usSat, 01 Nov 2025 00:00:00 +0000ROSE: A Reward-Oriented Data Selection Framework for LLM Task-Specific Instruction Tuninghttps://halfrost.me/publication/rose-reward-oriented-data-selection/Sat, 01 Nov 2025 00:00:00 +0000https://halfrost.me/publication/rose-reward-oriented-data-selection/