Research map / Large language models / Agents, code and math
Automated Formal Proofs of Combinatorial Identities via Wilf–Zeilberger Guidance and LLMs
Beibei Xiong, Hangyu Lv, Junqi Liu, Yisen Wang, Shaoshi Chen, Jianlin Wang, Zhengfeng Yang, Lihong Zhi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add WZ-Sketch decomposition to untrained Goedel-Prover-V2 (ablation) | Goedel-Prover-V2-8B · LCI-Test | proved (of 100): 9/100 -> 9/100 | No clear effect |
| replace Goedel-Prover-V2 by trained WZ-Prover for direct proving (ablation) | Goedel-Prover-V2-8B · LCI-Test | proved (of 100): 9/100 -> 12/100 | Improved |
| add WZ-Sketch guidance to WZ-Prover (full WZ-LLM) (ablation) | WZ-Prover · LCI-Test | proved (of 100): 12/100 -> 34/100 | Improved |
| replace Goedel-Prover-V2 by WZ-Prover as lemma prover (ablation) | WZ-Sketch · LCI-Test sketch subgoals | lemma accuracy (%): 47.88% -> 73.34% | Improved |
| replace Goedel-Prover-V2 by WZ-Prover as lemma prover (ablation) | WZ-Sketch · LCI-Test sketch problems | problems solved: 0 / 46 -> 29 / 46 | Improved |
| add WZ-Sketch to Goedel-Prover-V2 (ablation) | Goedel-Prover-V2-8B · CombiBench | proved (of 100): 12/100 -> 13/100 | No clear effect |
| add WZ-Sketch to Goedel-Prover-V2 (ablation) | Goedel-Prover-V2-8B · PutnamBench-Comb | proved (of 36): 0/36 -> 0/36 | No clear effect |
| replace Goedel-Prover-V2 by WZ-Prover (ablation) | Goedel-Prover-V2-8B · CombiBench | proved (of 100): 12/100 -> 15/100 | Improved |
| replace Goedel-Prover-V2 by WZ-Prover (ablation) | Goedel-Prover-V2-8B · PutnamBench-Comb | proved (of 36): 0/36 -> 3/36 | Improved |
| add sketch-guided route to WZ-Prover direct route (ablation) | WZ-Prover · CombiBench | proved (of 100): 15/100 -> 16/100 | Improved |
| add sketch-guided route to WZ-Prover direct route (ablation) | WZ-Prover · PutnamBench-Comb | proved (of 36): 2/36 -> 3/36 | Improved |
| add expert-verified iteration to seed SFT (ablation) | WZ-Prover · LCI-Test | proved (of 100): pass@32 9/100 -> 10/100 | Improved |
Closest papers by meaning
- ProofAug: Efficient Neural Theorem Proving via Fine-grained Proof Structure Analysis (ICML 2025 · similarity 0.86)
- APOLLO: Automated LLM and Lean Collaboration for Advanced Formal Reasoning (NeurIPS 2025 · similarity 0.86)
- 3D-Prover: Diversity Driven Theorem Proving With Determinantal Point Processes (NeurIPS 2025 · similarity 0.85)
- LEGO-Prover: Neural Theorem Proving with Growing Libraries (ICLR 2024 · similarity 0.85)
- HERMES: Towards Efficient and Verifiable Mathematical Reasoning in LLMs (ICML 2026 · similarity 0.85)
- MPS-Prover: Advancing Stepwise Theorem Proving by Multi-Perspective Search and Data Curation (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
