{"id":"arxiv/2611.11537","url":"https://arxiv.org/abs/2611.11537","tags":["quant-ph","cs.AI"],"score":0,"title":"Weight-Space Programming of Variational Quantum Circuits: Large Language Models Write Quantum Circuit Parameters Without an Optimizer","domain":"quant-ph","source":"arxiv","authors":["Debarshi Kundu","Swaroop Ghosh","Vasant G Honavar"],"reviews":0,"subject":"quant-ph","summary":"Variational quantum algorithms (VQAs) optimize circuit angles through repeated quantum-device evaluations, often facing vanishing gradients and high optimization costs. We ask whether a large language model (LLM) can write these angles directly. Extending *weight-space programming* from classical networks to quantum circuits, we provide a frontier LLM (\\gpt) with the problem, circuit, and gate conventions. Its angles are executed exactly as written, without returning losses, gradients, or accuracy feedback. Across 139 requests, we test handwritten-digit classification, VQE ground-state…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-11-08","reactions":{}}