Tasks
A catalog of multimodal tasks for evaluating modern LLMs.
The taxonomy indicates which skills of the model are tested by the test
Choose a modality
All tasks
Private
{{ name }}
Tested skills
All taxa
{{ skill.title }}
Expand the list of skills
Collapse the list of skills
Filters
Reset
Choose a modality
{{ name }}
Tested skills
{{ skill.title }}
The name of the set
Top Score | Human Baseline
Metric
For tasks with two or more metrics, the score is considered as the average of all metrics.
Tested skills
{{ task.scores.max?.toFixed(3) }}
TS
{{ task.scores.human?.toFixed(3) }}
HB
{{ metric.title }}
{{ skill.title }}
{{ skill.title }}
More {{ task.skills.length - 2 }}
mera@a-ai.ru
GitHub
hugging face