Social Bias Frames: Reasoning about Social and Power Implications of Language
Maarten Sap, Saadia Gabriel, Lianhui Qin, Dan Jurafsky, Noah A. Smith, Yejin Choi
摘要
Warning: this paper contains content that may be offensive or upsetting. Language has the power to reinforce stereotypes and project social biases onto others. At the core of the challenge is that it is rarely what is stated explicitly, but rather the implied meanings, that frame people's judgments about others. For example, given a statement that "we shouldn't lower our standards to hire more women," most listeners will infer the implicature intended by the speaker -that "women (candidates) are less qualified." Most semantic formalisms, to date, do not capture such pragmatic implications in which people express social biases and power differentials in language. We introduce SOCIAL BIAS FRAMES, a new conceptual formalism that aims to model the pragmatic frames in which people project social biases and stereotypes onto others. In addition, we introduce the Social Bias Inference Corpus to support large-scale modelling and evaluation with 150k structured annotations of social media posts, covering over 34k implications about a thousand demographic groups. We then establish baseline approaches that learn to recover SOCIAL BIAS FRAMES from unstructured text. We find that while stateof-the-art neural models are effective at highlevel categorization of whether a given statement projects unwanted social bias (80% F 1 ), they are not effective at spelling out more detailed explanations in terms of SOCIAL BIAS FRAMES. Our study motivates future work that combines structured pragmatic inference with commonsense reasoning on social implications. post off? int? lewd? group? targeted group implied statement ingroup? Kendrick just dropped the coldest video of all fucking time no no no What a fucking stupid bitch shut the fuck up already yes yes no need to fuck sum bad , I got some pressure built up :P no no yes I take a bitch home give her long dick yes yes yes no You annoyed me every day for months you're a fucking moron yes yes no no My problem with Korean artists: I dont know how to pronounce your name I can't hanggul to save my life yes no no yes Korean folks have weird names no I'll continue this thread till non-lifting fags stop lifting like idiots at the gym... Which is never. yes yes no yes gay men are weak and not manly no I thought drugs were the only things black people could shoot up Boy was I wrong yes yes no yes Black folks do drugs no kill people commit shootings
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper106
- GLaM: Efficient Scaling of Language Models with Mixture-of-ExpertsNan Du, Yanping Huang, Andrew M. Dai, Simon Tong 等ICML 2022 · 被引用 1,173 次
- Towards Understanding and Mitigating Social Biases in Language ModelsPaul Pu Liang, Chiyu Wu, Louis-Philippe Morency, Ruslan SalakhutdinovICML 2021 · 被引用 495 次
- Latent Hatred: A Benchmark for Understanding Implicit Hate SpeechMai ElSherief, Caleb Ziems, David Muchlinski, Vaishnavi Anupindi 等EMNLP 2021 · 被引用 159 次
- The Moral Integrity Corpus: A Benchmark for Ethical Dialogue SystemsCaleb Ziems, Jane A. Yu, Yi-Chia Wang, Alon Y. Halevy 等ACL 2022 · 被引用 127 次
- Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language ModelsMyra Cheng, Esin Durmus, Dan JurafskyACL 2023 · 被引用 89 次
它引用的顶会 Paper1
相关 Paper
- Multi-Dimensional Gender Bias ClassificationEmily Dinan, Angela Fan, Ledell Wu, Jason Weston 等EMNLP 2020 · 被引用 7 次
- Social-Group-Agnostic Bias Mitigation via the Stereotype Content ModelAli Omrani, Alireza Salkhordeh Ziabari, Charles Yu, Preni Golazizian 等ACL 2023 · 被引用 13 次
- Are Stereotypes Leading LLMs' Zero-Shot Stance Detection ?Anthony Dubreuil, Antoine Gourru, Christine Largeron, Amine TrabelsiEMNLP 2025 · 被引用 1 次
- Bi-directional Bias Attribution: Debiasing Large Language Models without Modifying PromptsYujie Lin, Kunquan Li, Yixuan Liao, Xiaoxin Chen 等ICLR 2026 · 被引用 6 次
- Social Story Frames: Contextual Reasoning about Narrative Intent and ReceptionJoel Mire, Maria Antoniak, Steven R. Wilson, Zexin Ma 等ACL 2026 · 被引用 2 次
