TACO is a task-aware LLM framework with abbreviation expansion, description generation, and task-based revision steps that improves downstream tabular NLP performance by up to 32%.
Ting Cai, Stephen Sheen, and AnHai Doan
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
-
TACO: Task-Aware Column Description Generation Using LLMs
TACO is a task-aware LLM framework with abbreviation expansion, description generation, and task-based revision steps that improves downstream tabular NLP performance by up to 32%.