WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks
Paper • 2406.13264 • Published
workflow_id stringlengths 23 25 | task_id int64 0 811 | intent stringlengths 13 248 | site stringclasses 4
values | start_url stringclasses 32
values | sop stringlengths 45 4.32k | trace stringlengths 143k 28.7M |
|---|---|---|---|---|---|---|
0 @ 2023-12-25-15-10-58 | 0 | What is the top-1 best-selling product in 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling product in 2022?\n\n1. Click on the \"Reports\" button on the far le(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2023-12-25T15:11:04.612219\", \"secs_(...TRUNCATED) |
0 @ 2023-12-25-16-43-12 | 0 | What is the top-1 best-selling product in 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling product in 2022\n1. On the left bar, click on the \"Reports\" button(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2023-12-25T16:43:17.762546\", \"secs_(...TRUNCATED) |
0 @ 2024-01-05-01-55-35 | 0 | What is the top-1 best-selling product in 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling product in 2022?\n1. Navigate to the sidebar located on the left sid(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-05T01:55:42.343495\", \"secs_(...TRUNCATED) |
0 @ 2024-01-08-14-57-29 | 0 | What is the top-1 best-selling product in 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling product in 2022\n1. Click on 'REPORTS' in the left sidebar.\n2. Clic(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-08T14:57:35.689251\", \"secs_(...TRUNCATED) |
0 @ 2024-01-13-04-37-18 | 0 | What is the top-1 best-selling product in 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling product in 2022\n1. Click on the “REPORTS” button in the dark gr(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-13T04:38:53.176195\", \"secs_(...TRUNCATED) |
1 @ 2023-12-25-15-44-04 | 1 | What is the top-1 best-selling brand in Quarter 1 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling brand in Quarter 1 2022?\n\n1. Click on the \"Reports\" button on th(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2023-12-25T15:44:12.727660\", \"secs_(...TRUNCATED) |
1 @ 2023-12-25-17-08-00 | 1 | What is the top-1 best-selling brand in Quarter 1 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling brand in Quarter 1 2022\n1. On the left bar, click on the \"Reports\(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2023-12-25T17:08:05.133783\", \"secs_(...TRUNCATED) |
1 @ 2024-01-05-01-57-42 | 1 | What is the top-1 best-selling brand in Quarter 1 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling brand in Quarter 1 2022\n1. Navigate to the sidebar located on the l(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-05T01:57:48.609040\", \"secs_(...TRUNCATED) |
1 @ 2024-01-08-15-20-51 | 1 | What is the top-1 best-selling brand in Quarter 1 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling brand in Quarter 1 2022\n1. Click on 'REPORTS' in the left sidebar.\(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-08T15:20:55.761475\", \"secs_(...TRUNCATED) |
1 @ 2024-01-13-04-50-53 | 1 | What is the top-1 best-selling brand in Quarter 1 2022 | shopping_admin | http://ec2-3-130-83-246.us-east-2.compute.amazonaws.com:7780/admin | "What is the top-1 best-selling brand in Quarter 1 2022\n1. Click on the “REPORTS” button in the(...TRUNCATED) | "[{\"type\": \"state\", \"data\": {\"id\": 0, \"timestamp\": \"2024-01-13T04:52:11.198578\", \"secs_(...TRUNCATED) |
zstd-compressed Parquet conversion of the JSON/Text traces from the WONDERBREAD dataset (Paper, Original Zenodo).
The original 7.4GB of text-only traces has been packed into a single Parquet file; The highly repetitive AXTrees and HTML DOM structures in the dataset compress extremely efficiently.
Data subsets:
default (viewer-safe): excludes the heavy trace JSON strings to prevent browser OOM on the Hugging Face dataset viewer.full: includes all columns, including the raw DOM snapshots and action history in the trace column.Each row represents a single workflow demonstration:
workflow_id (string): The original trace ID including the timestamp (e.g., 102 @ 2023-12-27-00-57-01)task_id (int32): The parsed integer WebArena task ID (e.g., 102)intent (string): The natural language intent of the task (e.g., What are the top-2 best-selling product in 2022)site (string): The WebArena site (e.g., shopping_admin)start_url (string): The starting URL of the workflowsop (string): The raw text of the Standard Operating Procedure (SOP - ...txt)trace (string, only in full config): The raw stringified JSON array containing the DOM states and actions.from datasets import load_dataset
import json
# Lightweight subset (default, viewer-safe)
ds = load_dataset("tkukurin/wonderbread")
train_split = ds["train"]
# Full dataset with the heavy JSON trace strings
ds_full = load_dataset("tkukurin/wonderbread", "full")
# Example: Parsing the JSON trace back into a list of steps
sample_trace_str = ds_full["train"][0]["trace"]
trace_list = json.loads(sample_trace_str)
@inproceedings{wornow2024wonderbread,
title={WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks},
author={Wornow, Michael and others},
booktitle={arXiv preprint arXiv:2406.13264},
year={2024}
}