BENCHCOMPASS: From Scores to Signals for Training and Harness Decisions in Payment-Domain LLMs
BENCHCOMPASS is introduced, a payment-domain benchmark whose construction pipeline builds scenario-grounded tasks from typed evidence packs, applies LLM-based quality checks, creates task-input attack variants, and reserves final item admission for domain experts.
Si-Jie Dong, Wei-Feng Ren, Xuan-Wei Hu et al.
· 0 citations