Dataset · callanwu
Bilingual questions and answers for evaluating web agents
callanwu/WebWalkerQA
A collection of 680 questions with answers and site navigation paths for testing AI agents.
What's inside
The dataset in JSON format contains 680 questions and answers from the WebWebWalker set in English and Chinese (metadata specifies a size of 10K–100K). Each example includes a starting URL, reference page transition path, number of steps, domain, difficulty level, and source site.
What you can do with it
- Test AI agents in multi-step information retrieval tasks on real sites
- Evaluate models' ability to plan navigation paths across web pages
- Compare web search quality of English and Chinese systems
- Tasks
- Question answering
- Languages
- Chinese, English
- Size
- 10K–100K rows
- Downloads this month
- 7.4K
- Likes
- 53
How to load it
from datasets import load_dataset
ds = load_dataset("callanwu/WebWalkerQA")