Team Ai
Datasetpublic

chenjianhui0428/HSCodeComp

HSCodeComp: A Realistic and Expert-Level Benchmark for Deep Search Agents in Hierarchical Rule Application Paper | Code | Dataset on Hugging Face ⭐ MarcoPolo Team ⭐ Alibaba Group 🗂️ Data 📌 Overview HSCodeComp is the first realistic, expert-level e-commerce benchmark designed to evaluate deep search agents on their ability to perform Level-3 knowledge—hierarchical rule application—a critical yet overlooked capability in current agent evaluation… See the full description on the dataset page: https://huggingface.co/datasets/chenjianhui0428/HSCodeComp.

sourceHugging Faceapache-2.0updated 3mo agoView on Hugging Face
0likes125downloads

Nothing at this path on main. The folder may be empty, or the revision may not exist.

chenjianhui0428/HSCodeComp · main · files are served by the source, never re-hosted here