The authors integrate an LLM judge into the semantics of a paraconsistent logic (AC) and show bilateral verification/refutation improves macro-F1 by about 6 points at the cost of abstaining on roughly half of the questions.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Sound and Complete Neurosymbolic Reasoning with LLM-Grounded Interpretations
The authors integrate an LLM judge into the semantics of a paraconsistent logic (AC) and show bilateral verification/refutation improves macro-F1 by about 6 points at the cost of abstaining on roughly half of the questions.