Jigsaw Unintended Bias in Toxicity Classification

By Q. Do. 2019

143 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

On the Intrinsic and Extrinsic Fairness Evaluation Metrics for Contextualized Language Representations (2022, unknown): “…the fairness of the language models (Guo and Caliskan, 2021; Ku-rita et al., 2019; Nadeem et al., 2020; Nangia et al., 2020), whereas extrinsic fairness metrics evaluate the fairness of the whole system through down-stream predictions (Dhamala et al., 2021; Jigsaw, 2019; De-Arteaga et al., 2019).”
Your fairness may vary: Pretrained language model fairness in toxic text classification (2021, background): ““Unintended Bias in Toxicity Classification” Kaggle competition (Jigsaw, 2019) that contains online comments on news articles, and 2) HateXplain, a dataset recently introduced with the intent of studying explanations for offensive and hate speech”
NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models (2024, unknown): “…(Casanueva et al., 2020), Emotion (Saravia et al., 2018), IMDB (Maas et al., 2011), MTOPDomain/MTOPIntent (Li et al., 2021), ToxicConversations (Adams et al., 2019), TweetSentimentExtraction (Maggie, 2020), AmazonPolarity (McAuley & Leskovec, 2013b), Mas-siveScenario/MassiveIntent (FitzGerald…”
What do Bias Measures Measure? (2021, methodology): “Along with this dataset, Jigsaw (2019) present a bias evaluation following that of Borkan et al. (2019) by comparing the AUC scores from different subgroups.”