Skip to content

Dataset · ArtificialAnalysis

Public sample for evaluating office AI agents

ArtificialAnalysis/AA-Briefcase-Lite

An open scenario from Artificial Analysis for testing AI agent performance on complex work tasks.

What's inside

This is a public demonstration scenario from the AA-Briefcase benchmark, including English documents and texts (under 1 thousand records). The original closed test consists of four scenarios simulating long-term specialist work on real work artifacts, including data analysis tasks. An open example of this workflow is published here.

What you can do with it

  • Exploring the autonomous AI agent evaluation format from Artificial Analysis
  • Testing agent pipelines on multi-step document tasks
Tasks
Other
Languages
English
Size
up to 1K rows
Downloads this month
5.7K
Likes
16

How to load it

from datasets import load_dataset

ds = load_dataset("ArtificialAnalysis/AA-Briefcase-Lite")