Practitioners trying to measure representational harms in LLM-based systems often cannot use public measurement instruments, either because the instruments lack validity, specificity, interpretability, or actionability, or because practical and institutional barriers block their uptake.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Understanding and Meeting Practitioner Needs When Measuring Representational Harms Caused by LLM-Based Systems
Practitioners trying to measure representational harms in LLM-based systems often cannot use public measurement instruments, either because the instruments lack validity, specificity, interpretability, or actionability, or because practical and institutional barriers block their uptake.