Search papers, labs, and topics across Lattice.
This paper introduces VirtualSet, an innovative interface that allows large language models (LLMs) to generate and interact with typed ontology worlds instead of relying on SQL for querying enterprise data. By implementing Generic Constraint Projection (GCP), VirtualSet ensures that invalid fields and actions are flagged as type errors before execution, significantly reducing the risk of hallucinated outputs. The results show that VirtualSet outperforms traditional SQL methods in accuracy while also preventing erroneous actions through a simulated decision-making process, achieving 67.5% accuracy compared to 63.5% for SQL with additional repair techniques.
VirtualSet not only boosts accuracy in LLM-generated queries but also eliminates the risk of executing hallucinated actions by enforcing type safety before execution.
Large language models increasingly read and act on enterprise data, but SQL gives a late error signal: hallucinated fields or relations can execute and return plausible wrong answers, while incorrect writes cannot be safely assessed after execution. We present VirtualSet, a live, receiver-typed ontology-world interface and generation target for LLMs. Instead of SQL, the model emits set expressions over entity-edge worlds. Generic Constraint Projection (GCP) checks expressions before execution, while future this preserves concrete receiver types through collection chains, turning invalid fields, edges, receivers, and actions into token-anchored type errors. Type-clean reads use a SQL fast path or bounded stream interpretation, with a parity oracle checking both paths over the exercised operator space. The same substrate supports guarded decisions: actions run first in a simulated world, and world-change events require external approval before actualization. On BIRD, we lift relational schemas into typed worlds and compare VirtualSet with direct SQL while holding the model, evidence, values, zero-shot setting, timeout, glossary, repair/voting, and grader constant where possible. On a frozen 1,072-question split, VirtualSet achieves 67.5% accuracy versus 63.5% for glossary-matched direct SQL with repair and voting (+4.0 points; McNemar exact p = 0.00117) using deepseek-reasoner. Full-corpus analysis finds no engine mis-computation of a type-clean expression; remaining errors arise from model semantics or gold defects. In a 30-body guard corpus, the write chain intercepts 20/20 hallucinated action bodies with zero false positives. VirtualSet thus remains competitive on SQL's home benchmark while providing pre-execution semantics for guarded decisions.
