A single LLM guided by a reusable prompt template matches or beats specialized models on the average of 15 oncology abstraction tasks, at roughly one-tenth the human effort.
Classifying cancer pathology reports with hierarchical self-attention networks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
Universal Abstraction: Harnessing Frontier Models to Structure Real-World Data at Scale
A single LLM guided by a reusable prompt template matches or beats specialized models on the average of 15 oncology abstraction tasks, at roughly one-tenth the human effort.