Rashomon ensembles of language models recover more complete human-annotated evidence for medical coding decisions than any single model.
Limitations of feature attribution in l ong text classification of standards
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
-
Complete Evidence Extraction with Model Ensembles: A Case Study on Medical Coding
Rashomon ensembles of language models recover more complete human-annotated evidence for medical coding decisions than any single model.