Granted Patent B2

US 12,650,991 · App. 19/213,030 · Granted Jun 9, 2026

Inventor: Samuel Bennett Cherny (Miami Beach, FL)

Assignee: Samuel Bennett Cherny

G06F16/24564

Abstract

Claims (105)

1 . A generalized natural language-operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output, the system comprising:

a plurality of delivery frameworks, each delivery framework included in the plurality of delivery frameworks operable to deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM, said LLM operating on one or more hardware processors, the plurality of delivery frameworks comprising:

a system framework, said system framework operable to deliver a system prompt structured in GNL-PA NLC code to the LLM, said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt the LLM operates as an operative system under one or more rules specified in the system prompt; and

a command framework, said command framework operable to deliver GNL-PA code to the LLM, said GNL-PA code for one or more commands executable within the operative system;

wherein the GNL-PA NLC is a formal language that operates as computer code, said formal language comprising a collection of letters taken from an alphabet and formed according to a set of grammar rules, said GNL-PA NLC increases meaning density of programming when the GNL-PA NLC is interpreted through the LLM to direct operations of the one or more hardware processors.

2 . The system of claim 1 wherein the system framework defines the system prompt.

3 . The system of claim 1 wherein the system framework is not visible by a user.

4 . A generalized natural language-operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output, the system comprising:

a plurality of delivery frameworks, each delivery framework included in the plurality of delivery frameworks operable to deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM, said LLM operating on one or more hardware processors, the plurality of delivery frameworks comprising:

a system framework, said system framework operable to deliver a system prompt structured in GNL-PA NLC code to the LLM, said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt the LLM operates as an operative system under one or more rules specified in the system prompt; and

a command framework, said command framework operable to deliver GNL-PA code to the LLM, said GNL-PA code for one or more commands executable within the operative system;

wherein:

each delivery framework included in the plurality of delivery frameworks comprises a plurality of components, each component included in the plurality of components comprises a plurality of segments, each segment included in the plurality of segments comprises a plurality of sections, each section included in the plurality of sections comprises a plurality of subsections;

each component, segment, section and subsection comprises one or more precepts;

each precept is a sequential constraint structured as a hierarchy; and

each precept is interpreted individually:

within the context of previous precepts in the hierarchy; and

affects interpretation of following precepts.

5 . A generalized natural language-operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output, the system comprising:

a plurality of delivery frameworks, each delivery framework included in the plurality of delivery frameworks operable to deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM, said LLM operating on one or more hardware processors, the plurality of delivery frameworks comprising:

a system framework, said system framework operable to deliver a system prompt structured in GNL-PA NLC code to the LLM, said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt the LLM operates as an operative system under one or more rules specified in the system prompt; and

a command framework, said command framework operable to deliver GNL-PA code to the LLM, said GNL-PA code for one or more commands executable within the operative system;

wherein the system framework comprises:

a system structure, said system structure being an initial element received at the LLM, said system structure comprising a text component describing subsequent elements to be received at the LLM and instructions for executing the subsequent elements;

a GNL-PA syntax and semantic instructions, said GNL-PA syntax and semantic instructions directing the LLM to accurately interpret the GNL-PA NLC;

a system architecture, the system architecture comprising the GNL-PA NLC, the GNL-PA NLC instantiates one or more protocols upon occurrence of a specified event included in a plurality of specified events, the system architecture constrains the LLM to operate the one or more protocols in a reproducible set of circumstances;

a module declaration, said module declaration translate the specified event, linked to the one or more protocols instantiated in response to the specified event, to one or more specific transformer blocks;

a command language architecture, said command language architecture linking each term, included in a plurality of terms, to a deterministic interpretation, included in a plurality of deterministic interpretations, said command language architecture setting each term, included in the plurality of terms, with a variable dynamic link to other terms included in the plurality of terms; and

a boot procedure, said boot procedure instructs the LLM to translate inputs input by a command structure, included in the command framework, into operational vector data.

6 . A generalized natural language-operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output, the system comprising:

a plurality of delivery frameworks, each delivery framework included in the plurality of delivery frameworks operable to deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM, said LLM operating on one or more hardware processors, the plurality of delivery frameworks comprising:

a system framework, said system framework operable to deliver a system prompt structured in GNL-PA NLC code to the LLM, said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt the LLM operates as an operative system under one or more rules specified in the system prompt; and

a command framework, said command framework operable to deliver GNL-PA code to the LLM, said GNL-PA code for one or more commands executable within the operative system, wherein the command framework comprises:

a command structure, said command structure comprising a text component that delineates the command framework to the LLM;

a natural language application architecture, said natural language application architecture defining a data structure to which a GNL-OS-based application conforms;

one or more directives, each of said one or more directives comprising operational instructions interpretable by the LLM; and

a finalized data structure, said finalized data structure comprising output of the LLM in response to the one or more directives.

7 . The system of claim 6 wherein the command framework delivers the following components of the command structure in a sequential order: the command structure, the natural language application architecture, the one or more directives and the finalized data structure.

8 . The system of claim 6 wherein:

the command framework sequentially processes the one or more directives; and

the finalized data structure is updated upon processing each directive.

9 . A method for constraining a large language model (“LLM”) to provide deterministic output, said LLM operating on one or more hardware processors, the method comprising:

delivering, from a system framework to the LLM, a system prompt structured in generalized natural language-program architecture natural language code (“GNL-PA NLC”), said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt, the LLM operates as an operative system under one or more rules specified in the system prompt; and

delivering, from a command framework to the LLM, GNL-PA code, said GNL-PA code for one or more commands executable within the operative system;

wherein the GNL-PA NLC is a formal language that operates as computer code, said formal language comprising a collection of letters retrieved from an alphabet and formed according to a set of grammar rules, said GNL-PA NLC increasing meaning density of programming when the GNL-PA NLC is interpreted through the LLM to direct operations of the one or more hardware processors.

10 . The method of claim 9 wherein the system framework defines the system prompt.

11 . The method of claim 9 wherein the system framework is not visible by a user.

12 . A method for constraining a large language model (“LLM”) to provide deterministic output, said LLM operating on one or more hardware processors, the method comprising:

delivering, from a system framework to the LLM, a system prompt structured in generalized natural language-program architecture natural language code (“GNL-PA NLC”), said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt, the LLM operates as an operative system under one or more rules specified in the system prompt; and

delivering, from a command framework to the LLM, GNL-PA code, said GNL-PA code for one or more commands executable within the operative system;

wherein:

the system framework includes a plurality of components, each component included in the system framework comprises a plurality of segments, each segment included in the plurality of segments comprises a plurality of sections, each section included in the plurality of sections comprises a plurality of subsections;

each component, segment, section and subsection comprises one or more precepts;

each precept is a sequential constraint structured as a hierarchy; and

each precept is interpreted individually:

within the context of previous precepts in the hierarchy; and

affects interpretation of following precepts.

13 . A method for constraining a large language model (“LLM”) to provide deterministic output, said LLM operating on one or more hardware processors, the method comprising:

delivering, from a system framework to the LLM, a system prompt structured in generalized natural language-program architecture natural language code (“GNL-PA NLC”), said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt, the LLM operates as an operative system under one or more rules specified in the system prompt; and

delivering, from a command framework to the LLM, GNL-PA code, said GNL-PA code for one or more commands executable within the operative system;

wherein:

the command framework includes a plurality of components, each component included in the command framework comprises a plurality of segments, each segment included in the plurality of segments comprises a plurality of sections, each section included in the plurality of sections comprises a plurality of subsections;

each component, segment, section and subsection comprises one or more precepts;

each precept is a sequential constraint structured as a hierarchy; and

each precept is interpreted individually:

within the context of previous precepts in the hierarchy; and

affects interpretation of following precepts.

14 . A method for constraining a large language model (“LLM”) to provide deterministic output, said LLM operating on one or more hardware processors, the method comprising:

delivering, from a system framework to the LLM, a system prompt structured in generalized natural language-program architecture natural language code (“GNL-PA NLC”), said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt, the LLM operates as an operative system under one or more rules specified in the system prompt; and

delivering, from a command framework to the LLM, GNL-PA code, said GNL-PA code for one or more commands executable within the operative system;

wherein the system prompt comprises:

a system structure, said system structure being an initial element received at the LLM, said system structure comprising a text component describing subsequent elements to be received at the LLM and instructions for executing the subsequent elements;

a GNL-PA syntax and semantic instructions, said GNL-PA syntax and semantic instructions directing the LLM to accurately interpret the GNL-PA NLC;

a system architecture, the system architecture comprising the GNL-PA NLC, the GNL-PA NLC instantiates one or more protocols upon occurrence of a specified event included in a plurality of specified events, the system architecture constrains the LLM to operate the one or more protocols in a reproducible set of circumstances;

a module declaration, said module declaration translating the specified event, linked to the one or more protocols instantiated in response to the specified event, to one or more specific transformer blocks;

a command language architecture, said command language architecture linking each term, included in a plurality of terms, to a deterministic interpretation, included in a plurality of deterministic interpretations, said command language architecture setting each term, included in the plurality of terms, with a variable dynamic link to other terms included in the plurality of terms; and

a boot procedure, said boot procedure instructs the LLM to translate inputs input by a command structure, included in the command framework, into operational vector data.

15 . A method for constraining a large language model (“LLM”) to provide deterministic output, said LLM operating on one or more hardware processors, the method comprising:

delivering, from a system framework to the LLM, a system prompt structured in generalized natural language-program architecture natural language code (“GNL-PA NLC”), said system prompt constraining the LLM to operate as an operative system, wherein upon receiving the system prompt, the LLM operates as an operative system under one or more rules specified in the system prompt; and

delivering, from a command framework to the LLM, GNL-PA code, said GNL-PA code for one or more commands executable within the operative system;

wherein the command framework comprises:

a command structure, said command structure comprising a text component that delineates the command framework to the LLM;

a natural language application architecture, said natural language application architecture defining a data structure to which a GNL-OS-based application conforms;

one or more directives, each of said one or more directives comprising operational instructions interpretable by the LLM; and

a finalized data structure, said finalized data structure comprising output of the LLM in response to the one or more directives.

16 . The system of claim 15 wherein the command framework delivers the following components of the command structure in a sequential order: the command structure, the natural language application architecture, the one or more directives and the finalized data structure.

17 . The system of claim 15 wherein:

the command framework sequentially processes the one or more directives; and

the finalized data structure is updated upon processing each directive.

18 . An artificial intelligence (“AI”) interface operating on a processor in combination with a memory, the AI interface comprising:

an application programming interface (“API”), said API in electronic communication with a large language model (“LLM”);

a defined program architecture;

a set of directives;

a finalized data structure, the finalized data structure is updated at upon conclusion of execution of each directive included in the set of directives; and

a set of input format benchmarks for an input expected within the execution of each directive; and

a set of output format benchmarks for an output expected within the execution of each directive;

wherein:

the set of directives are configured to be executed sequentially; and

a single execution of the AI interface is executed by a command, said single execution of the AI interface instantiating the API to electronically communicate with the LLM, said command being an input created from a sequentially appended contents of:

a command structure;

the defined program architecture;

a selected directive from the set of directives being currently executed;

the finalized data structure;

the set of input format benchmarks; and

the set of output format benchmarks.

Continuity (2)

References Cited (11)