Comparing Llama-2 and GPT-3 LLMs for HPC kernels generation
Explore the source record for details and available documents.
Valero Lara, Pedro↗
SEARCH · Engineering Papers
Results for “GPT”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
NASA-GPT: Retrieval-Augmented Generation on NASA Technical Reports
Explore the source record for details and available documents.
SLS↗
Using Generative AI to implement the discrepancy checker for a Nearly Autonomous Management and Control System for Advanced Reactors
Developments related to generative artificial intelligence (AI) have brought a major breakthrough in AI. These developments are rapidly accelerating developments in different science and engineering applications. Nearly Autonomous Management and Control (NAMAC) system provides recommendations to the operator for maintaining the safety and performance of the reactor. The discrepancy checker (DC) is an important component of the NAMAC) system, whose goal is to determine if the plant is moving towards the expected system state after the control actions are injected. In this work, we explore generative AI methods, particularly, a generative pretrained transformer (GPT) for implementing the DC function in NAMAC. The GPT-based DC aims to alert the operator in situations outside NAMAC’s scope and act as a chatbot the operator can use to retrieve relevant information. This study involves two versions of GPT developed by OpenAI: GPT-3.5 and GPT-4. These GPTs are trained on huge amounts of undisclosed general domain datasets. We explored two methods to adapt GPTs for DC implementation in NAMAC: fine-tuning and retrieval augmented generation. A small knowledge base (information file) that encompasses rules for DC implementation and some general information related to NAMAC has been created to support DC implementation using GPT. In this work, the GPT-based DC implementations have been tested for their reasoning abilities, comprehension, information retrieval, and extraction abilities. It should be noted that this paper only presents a preliminary study to test the feasibility of DC implementation using generative AI technology. Given the potential risks and severe consequences associated with nuclear reactor applications, combined with the black-box nature of AI, extensive offline and online testing and reliability analyses of GPT-based DCs are needed for further developing such capabilities.
22 GENERAL STUDIES OF NUCLEAR REACTORS↗
Development of Adaptive Model Refinement (AMoR) for Multiphysics and Multifidelity Problems
This project investigated the development and utilization of Adaptive Model Refinement (AMoR) for nuclear systems simulation applications. AMoR refers to utilization of several models of physical phenomena which differ in prediction fidelity. If the highest fidelity model is judged to always provide or exceeded the desired fidelity, than if one can determine the difference in a Quantity of Interest (QoI) between the highest fidelity model and lower fidelity models, one could utilize the fidelity model that would just provide the magnitude of the QoI desired. Assuming lower fidelity models require less computational resources, in this manner computational efficiency can be realized provided the QoI value can be accurately and efficiently evaluated. This work utilized Generalized Perturbation Theory (GPT) to evaluate the QoI, by convoluting the GPT solution with the residual of the highest fidelity model determined using the solution from lower fidelity models. Specifically, a reactor core neutronics problem and thermal-hydraulics problem were studied to develop and utilize AMoR. The highest fidelity neutronics model was based upon the 3D space-time, two-group, nodal diffusion equations as solved in the NESTLE computer code. Added to the NESTLE code was the ability to determine the time-dependent GPT neutron flux. The lower fidelity neutronics model was based upon the point kinetics equations along with utilization of a prolongation operator to determine the 3D space-time, two-group flux. The highest fidelity thermal-hydraulics model was based upon the space-time equations governing fluid flow in a closed channel around a heat generating fuel rod. The Homogenous Equilibrium Mixture (HEM) model was used for the fluid and Finite Difference Method was applied to both the coolant and fuel pin energy conservation equations. The lower fidelity thermal-hydraulic model was based upon the same equations as used for the highest fidelity model but now with coarse spatial meshing, corrected somewhat by employing effective fuel heat conduction values. The effectiveness of switching between the highest fidelity model and lower fidelity model as a function of time was assessed using the neutronics problem. Based upon work completed to date, one concludes that the time switching is effective in annealing out differences between the highest and lower fidelity solutions. The effectiveness of using a lower fidelity GPT solution, along with a prolongation operator, to estimate the QoI was also assessed. The utilization of a lower fidelity GPT solution was done in an attempt to avoid the high computational burden associated with solving for the highest fidelity GPT solution. Based upon work completed to date, one concludes that the lower fidelity adjoint solution is not sufficiently accurate with regard to estimating the QoI; however, a formulation has been revealed that may provide a path for addressing this shortcoming.
21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
Local volume effects in the generalized pseudopotential theory
The generalized pseudopotential theory (GPT) is a compelling approach for deriving real-space transferable interatomic potentials. Using a coarse-grained electronic structure, one can explicitly determine the pair ion-ion and multi-ion interactions in simple and transition metals. While successful in determining bulk properties, in central force metals the GPT fails to describe crystal defects for which there is a significant local volume change. A previous paper [J. A. Moriarty and R. Phillips, Phys. Rev. Lett. 66, 3036 (1991)] found that by allowing the GPT total energy to depend upon some spatially averaged local electron density, the energetics of vacancies and surfaces could be calculated within experimental ranges. In this paper, we develop the formalism further by explicitly calculating the forces and stress tensor associated with this total energy. We call this scheme the adaptive GPT (aGPT) and it is capable of both molecular dynamics (MD) and molecular statics. We apply the aGPT to vacancy formation, divacancy binding, and stacking faults in hcp Mg. We also calculate the local electron density corrections to the bulk elastic constants and phonon dispersion for which there is refinement over the baseline GPT treatment. Furthermore, we demonstrate aGPT-MD simulation through the calculation of thermal expansion in magnesium to 700 K.
75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
A Risk-Informed Approach to Trustworthiness Assessment in Digital Twins-Based Autonomous Control
In autonomous control systems, digital twins (DTs) are used to perform diagnostic and prognostic functions. The trustworthiness of these DTs is dependent on quality and coverage of the training data, model accuracy and integrity of sensor data. This work introduces a methodology to determine the trustworthiness of a DT system given faulty sensor data using a risk informed approach. Bayesian Belief Networks (BBNs) are used to propagate uncertainties and determine the probability of trustable recommendations. The decision to trust the control action provided by the DT is based on the DT output, expert opinion, and severity of problems. The performance of DTs is reliant on the data they are trained on. When they encounter out of distribution data, the trustworthiness of the recommendations decreases. To address this issue, we include an expert component that provides input on sensor degradation. For this, we utilize a generative artificial intelligence (AI) model, such as Generative Pretrained Transformer (GPT). The GPT functions as an expert with broad knowledge. The GPT is fine-tuned to understand and discriminate sensor degradation scenarios using manufactured data. This methodology is demonstrated through a case study on a Nearly Autonomous Management and Control System (NAMAC) during a steady state scenario. Various sensor degradation types with different severity levels are considered. Degraded sensor data is processed by the DT system and the fine-tuned GPT. Finally, using the BBN, we combine the GPT information and the DT output with its sources of uncertainty. This provides an output regarding the trustworthiness of the DT recommendation.
22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
Consistent performance of large language models in rare disease diagnosis across ten languages and 4917 cases
Background Large language models (LLMs) are increasingly used medicine for diverse applications including differential diagnostic support. The training data used to create LLMs such as the Generative Pretrained Transformer (GPT) predominantly consist of English-language texts, but LLMs could be used across the globe to support diagnostics if language barriers could be overcome. Initial pilot studies on the utility of LLMs for differential diagnosis in languages other than English have shown promise, but a large-scale assessment on the relative performance of these models in a variety of European and non-European languages on a comprehensive corpus of challenging rare-disease cases is lacking. Methods We created 4917 clinical vignettes using structured data captured with Human Phenotype Ontology (HPO) terms with the Global Alliance for Genomics and Health (GA4GH) Phenopacket Schema. These clinical vignettes span a total of 360 distinct genetic diseases with 2525 associated phenotypic features. We used translations of the Human Phenotype Ontology together with language-specific templates to generate prompts in English, Chinese, Czech, Dutch, French, German, Italian, Japanese, Spanish, and Turkish. We applied GPT-4o, version gpt-4o-2024-08-06, and the medically fine-tuned Meditron3-70B to the task of delivering a ranked differential diagnosis using a zero-shot prompt. An ontology-based approach with the Mondo disease ontology was used to map synonyms and to map disease subtypes to clinical diagnoses in order to automate evaluation of LLM responses. Findings For English, GPT-4o placed the correct diagnosis at the first rank 19.9% and within the top-3 ranks 27.0% of the time. In comparison, for the nine non-English languages tested here the correct diagnosis was placed at rank 1 between 16.9% and 20.6%, within top-3 between 25.4% and 28.6% of cases. The Meditron3 model placed the correct diagnosis within the first 3 ranks for 20.9% of cases in English and between 19.9% and 24.0% for the other nine languages. Interpretation The differential diagnostic performance of LLMs across a comprehensive corpus of rare-disease cases was largely consistent across the ten languages tested. This suggests that the utility of LLMs in clinical settings may extend to non-English clinical settings.
Artificial intelligence↗
GlcNAc-1-P-transferase–tunicamycin complex structure reveals basis for inhibition of N-glycosylation
N-linked glycosylation is a predominant post-translational modification of protein in eukaryotes, and its dysregulation is the etiology of several human disorders. The enzyme UDP-N-acetylglucosamine:dolichyl-phosphate N-acetylglucosaminephosphotransferase (GlcNAc-1-P-transferase or GPT) catalyzes the first and committed step of N-linked glycosylation in the endoplasmic reticulum membrane, and it is the target of the natural product tunicamycin. Tunicamycin has potent antibacterial activity, inhibiting the bacterial cell wall synthesis enzyme MraY, but its usefulness as an antibiotic is limited by off-target inhibition of human GPT. Our understanding of how tunicamycin inhibits N-linked glycosylation and efforts to selectively target MraY are hampered by a lack of structural information. Here we present crystal structures of human GPT in complex with tunicamycin. In conclusion, structural and functional analyses reveal the difference between GPT and MraY in their mechanisms of inhibition by tunicamycin. Here, we demonstrate that this difference could be exploited to design MraY-specific inhibitors as potential antibiotics.
59 BASIC BIOLOGICAL SCIENCES↗
AI in Science Communication
Generative AI has brought innovations across multiple fields, offering great tools for enhanced communication and efficiency. This project focused on developing a custom AI chatbot using OpenAI's Chat GPT (GPT-4o) to support the Fermilab communications team. An analysis identified Chat GPT as the optimal choice, leading to the adoption of its team version and the implementation of a real-time JSON schema for website scanning. Four distinct personas were created to tailor responses to specific audiences, and Fermilab's published content was uploaded to ensure tone consistency. The training involved iterative prompt trials, resulting in a responsive and effective communication assistant. Initial evaluations indicate that the custom GPT shows promise.
Valle, Diego↗
Implementation of Perturbation Theory and Sensitivity Capabilities in Griffin
Griffin is a Multiphysics Object-Oriented Simulation Environment (MOOSE) based reactor Multiphysics analysis application, jointly developed by Argonne and Idaho National Laboratories under the DOE-NE NEAMS program. This fiscal year, capabilities for reactivity and sensitivity evaluation using perturbation methods were implemented and verified. The First Order Perturbation Method (FOPT) was employed to compute reactivity worth resulting from small perturbations in input parameters, while the Generalized Perturbation Theory (GPT) was used to evaluate sensitivities of a range of response types, including reaction rate ratio, k-eigenvalue, neutron generation time, and effective delayed neutron fraction. These perturbation methods enable users to quantify how response quantities change due to a perturbation in a input parameter without explicitly performing an additional transport simulation for each perturbed state. In particular, the GPT formulation accounts for indirect effects arising from flux changes by solving generalized inhomogeneous equations, for which a Neumann series-based iterative solution method was developed and implemented in Griffin. The implemented reactivity and sensitivity evaluation capabilities were verified using two test problems: an infinite homogeneous system and a two-dimensional hexagonal core. The results showed excellent agreement with reference solutions obtained by a direct method based on finite difference approximation as well as GPT-based results from the PERSENT code, confirming the accuracy of both reactivity and sensitivity evaluations. Additionally, preliminary uncertainty quantification (UQ) results were obtained by combining the sensitivity values computed using GPT and external covariance data, demonstrating that the implemented sensitivity results can be reliably used for uncertainty calculations. To further demonstrate the generality and practical strength of the implementation, the sensitivity evaluation capability was successfully applied to the Empire microreactor with a geometrically complex design that poses significant modeling challenges. The results confirm that Griffin enables sensitivity evaluations even for irregular and highly heterogeneous reactor configurations, thereby establishing a foundation for UQ applications in advanced reactor designs and analyses.
22 GENERAL STUDIES OF NUCLEAR REACTORS↗
The AI-Base NDA Catalog
The project objective was to create a catalog of NDA capabilities so that as future challenges emerge, potential solutions can be quickly identified. The initial direction was to create a custom database where we would manually add various NDA technologies. During the year, LANL leadership heavily embraced AI tools and established LANL’s Enterprise ChatGPT license. To align with this vision, instead of a database we created a custom GPT to achieve the same capability. We provided a series of references on NDA technologies, the most comprehensive of which is the 2024 “PANDA manual” which is a 700-page textbook. A custom GPT is a semi-isolated version of OpenAI’s GPT model that can be tailored to specific tasks with context documents and instructions. Anyone with a LANL ChatGPT Enterprise account can access the NDA Catalog. Since the GPT exists on the green network, only non-sensitive questions may be asked. The Catalog can be used by simply asking questions in conversational English. We have found the Catalog to be quite accurate, even for heavily technical, complex queries. The Catalog allows the user to describe a measurement problem and it will respond with potential technology solutions.
96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
U.S. Wind Siting Regulation and Zoning Ordinances (2025)
A machine readable collection of documented wind siting ordinances at the state and local (e.g., county, township) level throughout the United States. The data were compiled using the Infrastructure Continuous Ordinance Mapping for Planning and Siting Systems (INFRA-COMPASS) tool, which leverages Large Language Models (LLMs) to automate the collection of local codes and ordinances applicable to energy infrastructure. URLs for the ordinance source documents are included in the Wind Ordinances spreadsheet. The GeoPackage file included below contains the jurisdiction shapes for each ordinance. Note that the GeoPackage file is formatted for ingestion by NREL's reVX setbacks tool and therefore does not contain any of the state-level regulations. NOTE: This data was collected with the help of generative AI. The Large Language Models used for this effort make mistakes. Always validate the data for critical use cases. This data is an update to a previously developed database of wind ordinances found in OEDI Submission 5733: see the "U.S. Wind Siting Regulation and Zoning Ordinances 2022" link below. INFRA-COMPASS version used for collection: v0.8.2 LLMs used for collection: GPT-4.1, GPT-4.1 mini, GPT-4.1 nano, GPT-4o mini
17 WIND ENERGY↗
A Generalized Perturbation Theory Solver In Rattlesnake Based On PETSc With Application To TREAT Steady State Uncertainty Quantification
Rattlesnake and MAMMOTH are the designated TREAT analysis tools currently being developed at the Idaho National Laboratory. Concurrent with development of the multi-physics, multi-scale capabilities, sensitivity analysis and uncertainty quantification (SA/UQ) capabilities are required for predicitive modeling of the TREAT reactor. For steady-state SA/UQ, that is essential for setting initial conditions for the transients, generalized perturbation theory (GPT) will be used. This work describes the implementation of a PETSc based solver for the generalized adjoint equations that constitute a inhomogeneous, rank deficient problem. The standard approach is to use an outer iteration strategy with repeated removal of the fundamental mode contamination. The described GPT algorithm directly solves the GPT equations without the need of an outer iteration procedure by using Krylov subspaces that are orthogonal to the operator’s nullspace. Three test problems are solved and provide sufficient verification for the Rattlesnake’s GPT capability. We conclude with a preliminary example evaluating the impact of the Boron distribution in the TREAT reactor using perturbation theory.
97 MATHEMATICS AND COMPUTING↗
AI in Science Communication
Generative AI has brought great innovations across multiple fields, offering great tools for enhanced communication and efficiency. This project focused on developing a custom AI chatbot using OpenAI's Chat GPT (GPT-4o) to support the Fermilab communications team. An analysis identified Chat GPT as the optimal choice, leading to the adoption of its team version and the implementation of a real-time JSON schema for website scanning. Four distinct personas were created to tailor responses to specific audiences, and Fermilab's published content was uploaded to ensure tone consistency. The training involved iterative prompt trials, resulting in a responsive and effective communication assistant. Initial evaluations indicate that the custom GPT shows promise.
Valle, Diego↗
An Introduction to Word Embeddings and Language Models
Language models have advanced at a phenomenal pace over the past decade. This document provides a short introduction to terminology, word embeddings (aka low-dimensional representations), and popular large-scale language models (LMs). Word embeddings are used to represent words as numerical vectors and are context-independent, meaning a word can only have a single representation (e.g., club can only be club sandwich, not golf club ). Language models can determine the probability of a given sequence of words occurring in a sentence and can provide context to distinguish between words and phrases that sound similar. LMs are context-dependent (e.g., club can be club sandwich or golf club ) and largely fall in two main classes – autoregressive and autoencoding models. Autoregressive models are pretrained on the classic language modeling task: guess the next token having read all the previous ones. Those models can be fine-tuned and achieve great results on many tasks, the most natural application is text generation. A typical example of such models is GPT, but others include GPT-2, GPT-3, CTLR, TRANSFORMER-XL, REFORMER, XLNET. Autoencoding models are pretrained by corrupting the input tokens in some way and trying to reconstruct the original sentence. They can be fine-tuned and achieve great results on many tasks such as text generation, but their most natural application is sentence classification or token classification. A typical example of such models is BERT, but others include ROBERTA, ALBERT, XML, XML-ROBERTA, FLAUBERT AND LONGFORMER.
97 MATHEMATICS AND COMPUTING↗
Language models for materials discovery and sustainability: Progress, challenges, and opportunities
Significant advancements have been made in one of the most critical branches of artificial intelligence: natural language processing (NLP). These advancements are exemplified by the remarkable success of OpenAI’s GPT-3.5/4 and the recent release of GPT-4.5, which have sparked a global surge of interest akin to an NLP gold rush. Here, in this article, we offer our perspective on the development and application of NLP and large language models (LLMs) in materials science. We begin by presenting an overview of recent advancements in NLP within the broader scientific landscape, with a particular focus on their relevance to materials science. Next, we examine how NLP can facilitate the understanding and design of novel materials and its potential integration with other methodologies. To highlight key challenges and opportunities, we delve into three specific topics: (i) the limitations of LLMs and their implications for materials science applications, (ii) the creation of a fully automated materials discovery pipeline, and (iii) the potential of GPT-like tools to synthesize existing knowledge and aid in the design of sustainable materials.
36 MATERIALS SCIENCE↗
ChemGraph as an agentic framework for computational chemistry workflows
Atomistic simulations are essential in chemistry and materials science but remain challenging to run due to the expert knowledge required for the setup, execution, and validation stages of these calculations. We present ChemGraph, an agentic framework powered by artificial intelligence and state-of-the-art simulation tools to streamline and automate computational chemistry and materials science workflows. ChemGraph leverages graph neural network-based foundation models for accurate yet computationally efficient calculations and large language models (LLMs) for natural language understanding, task planning, and scientific reasoning to provide an intuitive and interactive interface. We evaluate ChemGraph across 13 benchmark tasks and demonstrate that smaller LLMs (GPT-4o-mini, Claude-3.5-haiku, Qwen-2.5-14B) perform well on simple workflows, while more complex tasks benefit from using larger models. Importantly, we show that decomposing complex tasks into smaller subtasks through a multi-agent framework enables GPT-4o to reach perfect accuracy and smaller LLMs to match or exceed single-agent GPT-4o's performance in these benchmarks.
Computational chemistry↗
The ballad of LLM agents: philosophical reasoning for chemistry
Large language models (LLMs) show remarkable potential for scientific reasoning but often produce unreliable or scientifically unactionable outputs when faced with multi-step logic, domain grounding, and interpretability challenges, especially in complex fields like chemistry and materials science. Here, we introduce a framework of philosophical reasoning agents, inspired by canonical thinkers such as Socrates, Descartes, Kant, and Hume, to guide LLM behavior via structured prompt engineering. These agents embody distinct reasoning paradigms (dialectical inquiry, deductive logic, rule-based judgment, and empirical validation) and are evaluated across multiple chemistry subdomains, physical, analytical, general, inorganic, and organic chemistry, using the ChemBench benchmark. Our agentic prompting approach yields substantial accuracy gains on open-ended numerical chemistry questions, with gains of +11.5 percentage points for GPT-4o with Hume, +4.5 percentage points for GPT-5 with Kant, and +21.8 percentage points for GPT-5.1 with Socrates at the strict 1% error threshold, relative to the corresponding base models. Beyond accuracy, we observe benchmark-level model–agent performance patterns, suggesting that different prompting styles interact differently with each base model. These findings demonstrate that embedding philosophy-of-science principles into multi-agent frameworks can improve and produce interpretable, adaptive, and domain-aligned scientific LLMs.
Harb, Hassan [Argonne National Laboratory (ANL), A↗