arxiv:2606.00188
Kai Xu
kai-xu
AI & ML interests
Evaluations, safety training, AI for science
Recent Activity
liked a dataset 2 days ago
SALT-NLP/SWE-chat upvoted a paper 19 days ago
Is Agent Code Less Maintainable Than Human Code? commentedon a paper about 2 months ago
PaintBench: Deterministic Evaluation of Precise Visual Editing