A new dataset of 90 natural-language NP-completeness reduction proofs is introduced and used to benchmark reasoning in LLMs.
Proof verification and the hardness of approximation problems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
The Karp Dataset
A new dataset of 90 natural-language NP-completeness reduction proofs is introduced and used to benchmark reasoning in LLMs.