WZ-LLM combines Wilf-Zeilberger guidance with LLMs to reach 34% success proving 100 classic combinatorial identities in Lean 4, outperforming baselines like DeepSeek-V3.
Complete Lean 4 Proof
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Automated Formal Proofs of Combinatorial Identities via Wilf-Zeilberger Guidance and LLMs
WZ-LLM combines Wilf-Zeilberger guidance with LLMs to reach 34% success proving 100 classic combinatorial identities in Lean 4, outperforming baselines like DeepSeek-V3.