HatePrototypes built from few examples enable cross-benchmark transfer and efficient detection for both explicit and implicit hate speech.
OPT is pre-trained for causal language modeling with a 50k-token vocabulary, while BERT is pre- trained for masked language modeling with a 29k- token vocabulary
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
HatePrototypes: Interpretable and Transferable Representations for Implicit and Explicit Hate Speech Detection
HatePrototypes built from few examples enable cross-benchmark transfer and efficient detection for both explicit and implicit hate speech.