Prior or Feedback? What an LLM Uses When Adapting Neural Operators
Abstract
Do LLM scientific agents rely only on their initial task context, or do they adapt their decisions in response to experimental feedback? We study this question in neural operator adaptation, where a large language model (LLM) selects fine-tuning configurations under a limited trial budget. Across transfers within and between partial differential equation (PDE) families, the LLM achieves lower held-out test nRMSE than random search and Bayesian optimisation in nearly every matched comparison. Endpoint performance alone cannot distinguish what happens, so we verify each attribution with controlled interventions. Before observing any validation score, the LLM's first configuration already ranks near the top of the corresponding random-search pool, indicating a useful initial bias. A complementary cold-start intervention shows that the selected base learning rate shifts with the PDE description. Once feedback becomes available, reassigning validation scores among evaluated configurations changes the next proposal in every case tested, whereas a value-preserving rewrite produces no comparable aggregate effect. These interventions establish that the LLM's decision-level actions respond to the given task and observed outcomes, showing that it combines a task-dependent prior with sensitivity to experimental feedback.
Related
- Neural network-based multipartite entanglement classification with prior guidance from quantum uncertainty relations
- What Do Event Generators Compute? Flux Functionals from Detector Operators
- Pretrain finite element method: A pretraining and warm-start framework for PDEs via physics-informed neural operators
- Lecture notes on Physics Informed Neural Networks, Neural Operators, and their applications
- A hierarchy of physics-embedded neural operators for structural dynamic surrogate modelingA hierarchy of physics-embedded neural operators for structural dynamic surrogate modeling
BibTeX
@misc{chan2026prior,
title = {{Prior or Feedback? What an LLM Uses When Adapting Neural Operators}},
author = {Julian Chan and Javier Mora Jimenez},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12325},
eprint = {2610.12325},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12325},
}
RIS
TY - UNPB
TI - Prior or Feedback? What an LLM Uses When Adapting Neural Operators
AU - Julian Chan
AU - Javier Mora Jimenez
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.12325
UR - https://arxiv.org/abs/2610.12325
AB - Do LLM scientific agents rely only on their initial task context, or do they adapt their decisions in response to experimental feedback? We study this question in neural operator adaptation, where a large language model (LLM) selects fine-tuning configurations under a limited trial budget. Across transfers within and between partial differential equation (PDE) families, the LLM achieves lower held-out test nRMSE than random search and Bayesian optimisation in nearly every matched comparison. Endpoint performance alone cannot distinguish what happens, so we verify each attribution with controlled interventions. Before observing any validation score, the LLM's first configuration already ranks near the top of the corresponding random-search pool, indicating a useful initial bias. A complementary cold-start intervention shows that the selected base learning rate shifts with the PDE description. Once feedback becomes available, reassigning validation scores among evaluated configurations changes the next proposal in every case tested, whereas a value-preserving rewrite produces no comparable aggregate effect. These interventions establish that the LLM's decision-level actions respond to the given task and observed outcomes, showing that it combines a task-dependent prior with sensitivity to experimental feedback.
ER -