paper-with-me

Papers

HONEST: Measuring Hurtful Sentence Completion in Language Models

2021-06-01 · NAACL 2021 4 · Debora Nozza, Federico Bianchi, Dirk Hovy

Language models have revolutionized the field of NLP. However, language models capture and proliferate hurtful stereotypes, especially in text generation. Our results show that 4.3{\%} of the time, language models complete a sentence with a hurtful word. These cases are not random, but follow language and gender-specific patterns. We propose a score to measure hurtful sentence completions in language models (HONEST). It uses a systematic template- and lexicon-based bias evaluation methodology for six languages. Our findings suggest that these models replicate and amplify deep-seated societal stereotypes about gender roles. Sentence completions refer to sexual promiscuity when the target is female in 9{\%} of the time, and in 4{\%} to homosexuality when the target is male. The results raise questions about the use of these models in production settings.

📄 PDF Abstract BibTeX

Code (1)

milanlproc/honest

Tasks

Hate Speech DetectionHurtful Sentence CompletionMultilingual NLPSentenceSentence CompletionText Generation

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
DistilBERT 설명 없음
RoBERTa 설명 없음
Residual Connection 설명 없음
Weight Decay 설명 없음
Attention Dropout Attention Dropout is a type of dropout used in attention-based architectures, where elements are randomly dropped out of the…
Linear Warmup With Linear Decay Linear Warmup With Linear Decay is a learning rate schedule in which we increase the learning rate linearly for $n$ updates and then linearly decay afterwards.
WordPiece 설명 없음

Similar Papers 제목 키워드 기반

Measuring Gender Bias in West Slavic Language Models

2023-04-12 · Sandra Martinková, Karolina Stańczak, Isabelle Augenstein

Pre-trained language models have been known to perpetuate biases from the underlying datasets to downstream tasks. However, these findings are predominantly based on monolingual language models for English, whereas there…

Language Modelling

Measuring Harmful Sentence Completion in Language Models for LGBTQIA+ Individuals

2022-05-01 · LTEDI (ACL) 2022 5 · Debora Nozza, Federico Bianchi, Anne Lauscher, Dirk Hovy

Current language technology is ubiquitous and directly influences individuals’ lives worldwide. Given the recent trend in AI on training and constantly releasing new and powerful large language models (LLMs), there is a …

SentenceSentence Completion

Measuring Harmful Representations in Scandinavian Language Models

2022-11-21 · Samia Touileb, Debora Nozza

Scandinavian countries are perceived as role-models when it comes to gender equality. With the advent of pre-trained language models and their widespread usage, we investigate to what extent gender-based harmful and toxi…

FairBelief - Assessing Harmful Beliefs in Language Models

2024-02-27 · Mattia Setzu, Marta Marchiori Manerba, Pasquale Minervini, Debora Nozza

Language Models (LMs) have been shown to inherit undesired biases that might hurt minorities and underrepresented groups if such systems were integrated into real-world applications without careful fairness auditing. Thi…

Fairness

Showing LLM-Generated Code Selectively Based on Confidence of LLMs

2024-10-04 · Jia Li, Yuqi Zhu, Yongmin Li, Ge Li 외

Large Language Models (LLMs) have shown impressive abilities in code generation, but they may generate erroneous programs. Reading a program takes ten times longer than writing it. Showing these erroneous programs to dev…

Code Generation