internlm/WildClawBench
Visual Question AnsweringImage Text To TextQuestion AnsweringEN, ZHmit
The internlm/WildClawBench dataset is a EN, ZH visual question answering resource from internlm at 2026. With 17.3K downloads and 62 likes, it is actively used by the community. It is released under the mit license and is a n<1K-scale dataset.
About internlm/WildClawBench
WildClawBench
Hard, practical, end-to-end evaluation for AI agents — in the wild.
WildClawBench is an agent benchmark that tests what actually matters: can an AI agent do real work, end-to-end, without hand-holding?
We drop...
Details
- Task
- Visual Question Answering, Image Text To Text, Question Answering
- Language
- EN, ZH
- Format
- Parquet
- Rows / instances
- N/A
- Size
- n<1K
- Creator
- internlm
- Year
- 2026
- License
- mit
- Downloads
- 17310
- Likes
- 62