Pith. sign in

REVIEW 11 cited by

DataSciBench: An LLM Agent Benchmark for Data Science

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2502.13897 v1 pith:SIFPEXAI submitted 2025-02-19 cs.CL cs.AIcs.LG

DataSciBench: An LLM Agent Benchmark for Data Science

classification cs.CL cs.AIcs.LG
keywords metricsmodelscodedatadatascibenchevaluationcomprehensiveground
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Model (LLM) capabilities in data science. Recent related benchmarks have primarily focused on single tasks, easily obtainable ground truth, and straightforward evaluation metrics, which limits the scope of tasks that can be evaluated. In contrast, DataSciBench is constructed based on a more comprehensive and curated collection of natural and challenging prompts for uncertain ground truth and evaluation metrics. We develop a semi-automated pipeline for generating ground truth (GT) and validating evaluation metrics. This pipeline utilizes and implements an LLM-based self-consistency and human verification strategy to produce accurate GT by leveraging collected prompts, predefined task types, and aggregate functions (metrics). Furthermore, we propose an innovative Task - Function - Code (TFC) framework to assess each code execution outcome based on precisely defined metrics and programmatic rules. Our experimental framework involves testing 6 API-based models, 8 open-source general models, and 9 open-source code generation models using the diverse set of prompts we have gathered. This approach aims to provide a more comprehensive and rigorous evaluation of LLMs in data science, revealing their strengths and weaknesses. Experimental results demonstrate that API-based models outperform open-sourced models on all metrics and Deepseek-Coder-33B-Instruct achieves the highest score among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 11 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Orak: A Foundational Benchmark for Training and Evaluating LLM Agents on Diverse Video Games

    cs.AI 2025-06 unverdicted novelty 7.0

    Orak is a foundational benchmark providing training data, interfaces, and evaluation tools for LLM agents across diverse video game genres.

  2. RExBench: Can coding agents autonomously implement AI research extensions?

    cs.CL 2025-06 unverdicted novelty 6.0

    RExBench is a new benchmark showing that LLM coding agents fail to autonomously implement most realistic research extensions to prior AI papers.

  3. Do LLM-Generated Skills Make Better AI Data Scientists? A Component Ablation Across Data-Science Workflows

    cs.AI 2026-07 accept novelty 5.0

    Across 56 tasks, 9 model configurations, and 10,584 runs, LLM-generated skill files provided no reliable performance improvement over task-only prompting for data-science workflows.

  4. AgenticDataBench: A Comprehensive Benchmark for Data Agents

    cs.DB 2026-07 unverdicted novelty 5.0

    AgenticDataBench is a new benchmark covering realistic data science tasks across 15 domains using extracted skills and LLM-generated workflows to evaluate data agents at fine granularity.

  5. GRACE-DS: a Guarded Reward-guided Agent Correction Environment in Data Science

    cs.CL 2026-06 unverdicted novelty 5.0

    GRACE-DS supplies metrics and a guarded sandbox for end-to-end evaluation of LLM AutoML agents on organization-specific tabular tasks, with flexible iterative interaction outperforming baselines on hidden-test quality...

  6. Text Analytics Evaluation Framework: A Case Study on LLMs and Social Media

    cs.CL 2026-05 unverdicted novelty 5.0

    Presents a new question-based evaluation framework for LLMs on aggregated social media text and reports that performance declines with input scale, task complexity, and numerical operations beyond 500 instances.

  7. ProfiliTable: Profiling-Driven Tabular Data Processing via Agentic Workflows

    cs.AI 2026-05 unverdicted novelty 5.0

    ProfiliTable is a multi-agent system with profiler, generator, and evaluator components that outperforms baselines on 18 tabular task types via dynamic profiling and closed-loop refinement.

  8. ProfiliTable: Profiling-Driven Tabular Data Processing via Agentic Workflows

    cs.AI 2026-05 unverdicted novelty 5.0

    ProfiliTable is a profiling-driven multi-agent system that builds semantic context through exploration and closed-loop refinement to produce more reliable tabular data transformations than prior LLM approaches.

  9. Business Utility of Large Language Models as Exploratory Data Analysis Agents

    cs.CY 2026-05 unverdicted novelty 5.0

    Evaluation of 15 LLM configurations across four conditions in a supply chain EDA benchmark finds most lack sufficient repeatability for autonomous deployment, with GPT-5.4 at extra-high reasoning effort scoring highes...

  10. DRAFT: Task Decoupled Latent Reasoning for Agent Safety

    cs.LG 2026-02 unverdicted novelty 5.0

    DRAFT decouples agent safety judgment into latent extraction and reasoning stages, raising average benchmark accuracy from 63.27% to 91.18%.

  11. From LLM Reasoning to Autonomous AI Agents: A Comprehensive Review

    cs.AI 2025-04 accept novelty 4.0

    A survey consolidating benchmarks, agent frameworks, real-world applications, and protocols for LLM-based autonomous agents into a proposed taxonomy with recommendations for future research.