Papers › HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection

HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection

18 Dec 2020arXiv:2012.10289archive 2025-07-28

Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, Animesh Mukherjee

Hate speech is a challenging issue plaguing the online social media. While better models for hate speech detection are continuously being developed, there is little research on the bias and interpretability aspects of hate speech. In this paper, we introduce HateXplain, the first benchmark hate speech dataset covering multiple aspects of the issue. Each post in our dataset is annotated from three different perspectives: the basic, commonly used 3-class classification (i.e., hate, offensive or normal), the target community (i.e., the community that has been the victim of hate speech/offensive speech in the post), and the rationales, i.e., the portions of the post on which their labelling decision (as hate, offensive or normal) is based. We utilize existing state-of-the-art models and observe that even models that perform very well in classification do not score high on explainability metrics like model plausibility and faithfulness. We also observe that models, which utilize the human rationales for training, perform better in reducing unintended bias towards target communities. We have made our code and dataset public at https://github.com/punyajoy/HateXplain

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

punyajoy/HateXplain officialmentioned in papermentioned on GitHubpytorch report
Onepierre/Hate_Speech_Detection mentioned on GitHubpytorch report
darsh10/HateXplain-Darsh mentioned on GitHubpytorch report
hate-alert/HateXplain mentioned on GitHubpytorch report
sayani-kundu/11711-HateXplain mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Hate Speech DetectionText Classification

Datasets

Introduced by this paper, per the archive.

HateXplain

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Hate Speech Detection HateXplain BERT-HateXplain [Attn] AUROC 0.851 #3 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT-HateXplain [Attn] Accuracy 0.698 #3 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT-HateXplain [Attn] Macro F1 0.687 #3 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT-HateXplain [LIME] AUROC 0.851 #4 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT-HateXplain [LIME] Macro F1 0.687 #4 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT [Attn] AUROC 0.843 #5 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT [Attn] Accuracy 0.69 #5 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BERT [Attn] Macro F1 0.674 #5 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN-HateXplain [Attn] AUROC 0.805 #6 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN-HateXplain [Attn] Macro F1 0.629 #6 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN-Attn [Attn] AUROC 0.795 #7 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN-Attn [Attn] Accuracy 0.621 #7 of 11 Archive leaderboard report
Hate Speech Detection HateXplain CNN-GRU [LIME] AUROC 0.793 #8 of 11 Archive leaderboard report
Hate Speech Detection HateXplain CNN-GRU [LIME] Accuracy 0.629 #8 of 11 Archive leaderboard report
Hate Speech Detection HateXplain CNN-GRU [LIME] Macro F1 0.614 #8 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN [LIME] AUROC 0.767 #9 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN [LIME] Accuracy 0.595 #9 of 11 Archive leaderboard report
Hate Speech Detection HateXplain BiRNN [LIME] Macro F1 0.575 #9 of 11 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdamAttentionAttention DropoutBERTDense ConnectionsDropoutInterpretabilityLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionResidual ConnectionSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections