--- type: predictions tags: - autotrain - evaluation datasets: - mathemakitten/winobias_antistereotype_test_cot_v2 eval_info: task: text_zero_shot_classification model: inverse-scaling/opt-6.7b_ev

Hugging Face2022-11-07 更新100

Corpus for Asymmetrical Pluralism and the Normative Power of AI in Asian Englishes

语言模型偏见评估

英语变体研究

The corpus provides full prompt–response pairs and metadata for the study’s systematic analysis of differential audibility and asymmetrical pluralism. It includes interactions with three commercial la

Figshare2026-02-17 更新40

iproskurina/bias-in-bios-llama-seed40

语言模型偏见评估

职业传记生成

--- dataset_info: features: - name: seed_len dtype: int64 - name: temperature dtype: float64 - name: seed dtype: string - name: human_full dtype: string - name: human_conti

Hugging Face2025-12-13 更新140

quantiles/crows_pairs

语言模型偏见评估

社会偏见检测

CrowS-Pairs是一个用于评估掩码语言模型中社会偏见的挑战数据集。它包含1508个英语句子对，每个句子对包括一个更具偏见的句子（sent_more）和一个较少偏见的句子（sent_less），用于衡量模型在生成或评分时是否表现出社会偏见。数据集覆盖了多种偏见类型，如种族、性别、社会经济地位、残疾、国籍、性取向、外貌、宗教和年龄。每个示例都标注了偏见类型（bias_type）和刻板/反刻板标签

Hugging Face2026-04-27 更新70

HolisticBias Descriptors

资源简介：

相关数据集