Geometry of Values: Task Vector Composition for Ethical Preference Alignment in Language Models ▲ 2 💬 1 2026-09-22 arXiv HF 原文 GitHub ★1 摘要 (无摘要)