SICI is a new seven-dimensional complexity measure that predicts LLM stance detection accuracy and reveals distinct error regimes (over-attribution, boundary instability, abstention) as complexity rises.
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
-
SICI: A Semantic-Pragmatic Complexity Index Reveals Regime Shifts in LLM Stance Detection
SICI is a new seven-dimensional complexity measure that predicts LLM stance detection accuracy and reveals distinct error regimes (over-attribution, boundary instability, abstention) as complexity rises.