LLM-guided semantic vtrees for SDD compilation: code and data

Companion artifact for: Alexis de Colnet and Stefan Szeider, "LLM-guided semantic vtrees for SDD compilation", 20th Conference on Neurosymbolic Learning and Reasoning (NeSy 2026), Proceedings of Machine Learning Research, vol. 284, 2026. The package contains the implementation, the prompt driving the LLM stage, the experimental data underlying every table and headline number in the paper, three validation experiments, and the raw LLM outputs across the model sweep. It also contains supp_proof.pdf, the supplementary proof document with the formal vtree definitions and the full proof of the paper's Lemma 1; this is the document the paper's phrase "the supplementary material" refers to. Code is under MIT; data, prompts, and raw LLM outputs under CC BY 4.0