LLMs can learn annotator-specific label-explanation behavior from human label variation via cross-annotator preference optimization, outperforming prompting and standard fine-tuning on two sentence-pair tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Human Label Variation as Stable Signal: Learning Annotator-Specific Explanation Behavior via Cross-Annotator Preference Optimization
LLMs can learn annotator-specific label-explanation behavior from human label variation via cross-annotator preference optimization, outperforming prompting and standard fine-tuning on two sentence-pair tasks.