Presents a new expert-curated dataset of multi-turn counterspeech dialogues in five languages targeting hate against seven groups, with span annotations linking to verified external knowledge for RAG applications.
The Risk of Racial Bias in Hate Speech Detection
11 Pith papers cite this work, alongside 771 external citations. Polarity classification is still indexing.
11
Pith papers citing it
771
external citations · OpenAlex
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 11roles
background 1polarities
background 1representative citing papers
Six state-of-the-art LLMs systematically prefer Standard American English over AAE continuations, and a training-free activation steering method reduces this bias 5-20x more than prompting while preserving fluency.
A Dutch BERT model encodes gender linearly by epoch 20 but does not dynamically update its representations when explicit female cues contradict learned stereotypical associations in short sentence templates.
Dialectal robustness and generation are dissociated in LLMs: benchmarks are driven by pretraining and SFT while alignment reshapes generation invisibly to benchmarks, and the method maximizing dialectal reward is least preferred by human evaluators.
LLMs generate adequate counterspeech for co-occurring hate and misinformation in 40% of cases, with a mixed knowledge strategy from fact-checkers and NGOs proving most effective after expert revision.
CoGate-LSTM adds prototype-guided cosine feature-space gating to a character-level BiLSTM with multi-source embeddings and focal loss, reaching 0.881 macro-F1 on Jigsaw toxic comments while using 7.3M parameters and outperforming fine-tuned BERT by 6.9 points on minority labels.
The authors provide a detailed taxonomy of 21 risks associated with language models, covering discrimination, information leaks, misinformation, malicious applications, interaction harms, and societal impacts like job loss and environmental costs.
Toxicity in language models is disproportionately encoded in early MLP layers and can be localized via activation differentials then suppressed at inference time without gradient descent.
Automated hate speech detectors show poor alignment with heterogeneous in-group judgments on reclaimed slur usage, driven by low inter-annotator agreement and contextual features like derogatory intent.
Fine-tuning DistilRoBERTa on subreddit corpora and ranking words by embedding cosine-similarity drop isolates community-specific slang in the bottom 10th percentile.
A survey that catalogs threat models, detection approaches, and mitigation strategies for toxicity in multilingual LLMs while identifying challenges such as uneven language coverage and culturally variable harm definitions.
citing papers explorer
-
CATCH-ME if you RAG: a dataset of Contextually Annotated multi-Turn Counterspeech against Hate and Misinformation Exchanges
Presents a new expert-curated dataset of multi-turn counterspeech dialogues in five languages targeting hate against seven groups, with span annotations linking to verified external knowledge for RAG applications.
-
LLMs Silently Correct African American English: Auditing and Mitigating Dialect Bias via Activation Steering
Six state-of-the-art LLMs systematically prefer Standard American English over AAE continuations, and a training-free activation steering method reduces this bias 5-20x more than prompting while preserving fluency.
-
Is She Even Relevant? When BERT Ignores Explicit Gender Cues
A Dutch BERT model encodes gender linearly by epoch 20 but does not dynamically update its representations when explicit female cues contradict learned stereotypical associations in short sentence templates.
-
DiaLLM: An Investigation into the Robustness-Generation Gap in English Dialect Adaptation
Dialectal robustness and generation are dissociated in LLMs: benchmarks are driven by pretraining and SFT while alignment reshapes generation invisibly to benchmarks, and the method maximizing dialectal reward is least preferred by human evaluators.
-
Assisted Counterspeech Writing at the Crossroads of Hate Speech and Misinformation
LLMs generate adequate counterspeech for co-occurring hate and misinformation in 40% of cases, with a mixed knowledge strategy from fact-checkers and NGOs proving most effective after expert revision.
-
CoGate-LSTM: Prototype-Guided Feature-Space Gating for Mitigating Gradient Dilution in Imbalanced Toxic Comment Classification
CoGate-LSTM adds prototype-guided cosine feature-space gating to a character-level BiLSTM with multi-source embeddings and focal loss, reaching 0.881 macro-F1 on Jigsaw toxic comments while using 7.3M parameters and outperforming fine-tuned BERT by 6.9 points on minority labels.
-
Ethical and social risks of harm from Language Models
The authors provide a detailed taxonomy of 21 risks associated with language models, covering discrimination, information leaks, misinformation, malicious applications, interaction harms, and societal impacts like job loss and environmental costs.
-
Where Does Toxicity Live? Mechanistic Localization and Targeted Suppression in Language Models
Toxicity in language models is disproportionately encoded in early MLP layers and can be localized via activation differentials then suppressed at inference time without gradient descent.
-
IYKYK (But AI Doesn't): Automated Content Moderation Does Not Capture Communities' Heterogeneous Attitudes Towards Reclaimed Language
Automated hate speech detectors show poor alignment with heterogeneous in-group judgments on reclaimed slur usage, driven by low inter-annotator agreement and contextual features like derogatory intent.
-
Community-Specific Slang and Entity Detection via Semantic Shift in Fine-Tuned Language Models
Fine-tuning DistilRoBERTa on subreddit corpora and ranking words by embedding cosine-similarity drop isolates community-specific slang in the bottom 10th percentile.
-
A Survey of Toxicity Detection and Mitigation Strategies for Multilingual Language Models
A survey that catalogs threat models, detection approaches, and mitigation strategies for toxicity in multilingual LLMs while identifying challenges such as uneven language coverage and culturally variable harm definitions.