CoFEE enforces cognitive behaviors in LLMs for feature discovery, yielding 15.2% higher success rates, 29% fewer features, and 53.3% lower costs than vanilla prompting.
An empirical analysis of feature engineering for predictive modeling,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CoFEE: Reasoning Control for LLM-Based Feature Discovery
CoFEE enforces cognitive behaviors in LLMs for feature discovery, yielding 15.2% higher success rates, 29% fewer features, and 53.3% lower costs than vanilla prompting.