DABStep

Emerging

1papers using it

2026first seen

The 'DABStep' dataset/benchmark contains a collection of data analysis tasks used to evaluate the effectiveness of Process Reward Models (PRMs) in supervising data analysis agents and detecting errors in their reasoning processes.

🔎 Find this dataset

Papers using DABStep (1)

Rewarding the Scientific Process: Process-Level Reward Modeling for Agentic Data Analysis2026