AI Paper Review

Review machine learning, systems-for-ML, and LLM papers at rigorous conference-review depth. Use whenever the user asks for 审稿, 评审, 论文review, paper review, NeurIPS/ICML/ICLR水平, 按顶会水平看, critique a paper, score a submission, audit rebuttal readiness, inspect a LaTeX/PDF draft, or find mathematical inconsistencies, reproducibility gaps, checklist issues, broken references, experimental weaknesses, figure-text mismatches, artifact problems, or writing problems. Especially use for PAT-style or reviewer-style strengths/weaknesses, section-by-section issue finding, actionable revision advice, central-claim support checks, top reviewer attacks, and structured review forms.