Xinlong Huang
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Beyond Binary Labels: Diagnostic Evaluation of LLM Safety with Prompt-Level Rubrics
According to the abstract excerpt, the authors introduce DiagSafe, a bilingual English-Chinese safety benchmark that replaces safe/unsafe labels with prompt-level rubrics, and report 95.34% agreement between LLM judges and human annotations.