PROBAST tool for risk of bias assessment.

NIAID Data Ecosystem2026-03-14 收录

模型偏差评估

数据链接：

https://figshare.com/articles/dataset/PROBAST_tool_for_risk_of_bias_assessment_/22061773 数据链接链接失效反馈

官方服务：

资源简介：

Source: The authors (2022). (XLSX)

创建时间：

2023-02-09

CALM

模型偏差评估

自然语言处理

CALM是一个综合评估语言模型偏差的多任务基准数据集，由宾夕法尼亚州立大学计算机科学与工程系的研究团队创建。该数据集整合了16个现有数据集，用于问题回答、情感分析和自然语言推理三个NLP任务，共包含78,400个提示，用于测量性别和种族偏差。CALM通过筛选224个模板，并结合350个代表不同社会群体的常用人名，生成提示，旨在解决语言模型在处理不同社会群体时的性能差异问题。

arXiv2024-01-24 更新380

Unraveling overoptimism and publication bias in ML-driven science - Data for Experiment 2 and 3

医疗机器学习

模型偏差评估

In this folder, problem_1_all_20_M1_CIU and problem_2_all_20_M1_CIU provides necessary data for the two problems in the Experiment 2.real_data_CI_Speech provides meta data for speech-based classificat

DataCite Commons2025-06-01 更新50

Leave-one-out cross validation and Bootstrap .632+ estimator errors per features set.

模型偏差评估

特征选择验证

Leave-one-out cross validation and Bootstrap .632+ estimator errors per features set.

NIAID Data Ecosystem40

Feiye1213/lima-qwen25-pairrm-preference

模型偏差评估

答案排序

--- dataset_info: features: - name: prompt dtype: string - name: chosen dtype: string - name: rejected dtype: string - name: all_answers sequence: string - name: pairrm_ran

Hugging Face2026-04-27 更新50

Model evaluation and comparison.

评分者信度

模型偏差评估

Pearson correlation coefficients are included for correlation between the mean posterior bias of a rater and that rater’s mean’s deviation from the grand mean across all raters for current year and th

Figshare2018-04-04 更新50

PROBAST tool for risk of bias assessment.

资源简介：

相关数据集