dataagentbench
DataAgentBench-data
DataAgentBench — Datasets
Mirror of the large dataset files for DataAgentBench, a benchmark for data agents. These files (PostgreSQL dumps, SQLite/DuckDB databases, and MongoDB BSON) are hosted here instead of Git LFS so they can be downloaded freely.
Paper: Can AI Agents Answer Your Data Questions? A Benchmark for Data Agents (arXiv:2603.20576)
GitHub: ucbepic/DataAgentBench
Usage
Clone the benchmark repo and run its downloader, which fetches every file into… See the full description on the dataset page: https://huggingface.co/datasets/ruiyingm/DataAgentBench-data.data-agent-benchmarks
LongHorizon Full Data-Agent Benchmarks
Companion data artifacts for five complete evaluation tracks:
DataSciBench full55 / 167 metric entries
DABStep full450
DABStep-Research full100
DSBench Modeling full74
LongDS full68 / 2,225 turns
The companion GitHub repository contains processed manifests, evaluation code,
historical API ReAct baseline code, download/preparation tools, and the frozen
source lock. artifact_manifest.json records every uploaded object's size,
SHA-256… See the full description on the dataset page: https://huggingface.co/datasets/noel7Y/data-agent-benchmarks.dataagentbench-derived-official54-altimate-prompt-shell
DataAgentBench-Derived Official54 Altimate Prompt Shell
This public dataset contains 54 DataAgentBench-derived task prompt-shell rows compiled for the Altimate-centered DAB sandbox runtime. It is a derived compile/export artifact, not the official raw DataAgentBench release.
It is intended as a portable task/prompt/manifest source for downstream Altimate teacher rollouts, SFT construction, or RL data compilation. It is not a completed rollout dataset: the rows do not contain… See the full description on the dataset page: https://huggingface.co/datasets/forseasons/dataagentbench-derived-official54-altimate-prompt-shell.