New dataset Slang-Q aims to improve LLM understanding of queer slang

Researchers have introduced Slang-Q, a new dataset designed to address the underrepresentation of queer slang in Natural Language Processing research. This dataset, built on a taxonomy of 118 queer terms, contains user-generated English sentences paired with slang terms and their definitions. An initial evaluation using Slang-Q explores how well current language models can understand and define this specific type of language under different prompting conditions. AI

IMPACT This dataset could lead to more inclusive and accurate language models, improving their understanding of diverse linguistic expressions.

RANK_REASON The cluster contains an academic paper detailing a new dataset and its initial evaluation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• arXiv
• English
• Hugging Face
• natural language processing
• Slang-Q

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →