Jul 2026
CausalDS: Benchmarking Causal Reasoning in Data-Science Agents
CausalDS, a benchmark for evaluating causal reasoning in agentic data-science workflows, is introduced, which jointly evaluates symbolic causal reasoning, data science, uncertainty quantification, abstention, and tool use/coding.
Andrej Leban, Yuekai Sun
· arXiv.org · 0 citations