John MacCormick

Beam Search, Self-Consistency, and the Limits of Inference-Time Scaling for Grammar-Constrained Text-to-SQL in Small Language Models

The constrained case of this"model size vs. inference compute"trade-off, in which the model outputs are constrained by a strict grammar at inference time, is examined, which demonstrates that the constrained trade-off behaves differently from the unconstrained trade-off.