Tag
This paper introduces 'Compile by Training', a method that compiles natural-language function descriptions into local neural programs using teacher model synthesis and fine-tuning, achieving 83.6% semantic accuracy on FuzzyBench.
A tweet endorsing ProgramAsWeights as a working replacement for regex, demonstrating an email validation function compiled from natural language.
Program-as-Weights (PAW) introduces a programming paradigm where a 4B compiler translates natural-language specifications into compact neural artifacts executable by a 0.6B interpreter, achieving performance comparable to 32B models with drastically lower memory and inference cost.