Custom model instructions with language models

Patent metadata

Field	Value
Publication number	US-12430518-B2
Application number	US-202519065776-A
Country	US
Kind code	B2
Filing date	Feb 27, 2025
Priority date	Feb 27, 2024
Publication date	Sep 30, 2025
Grant date	Sep 30, 2025

Abstract

Disclosed herein are methods, systems, and computer-readable media for interacting with a language model using custom instructions. In one embodiment a method includes receiving, through an interface, custom instructions, the custom instructions comprising at least one of personal information or a response type preference, storing the custom instructions temporarily within a session specific cache, in response to a trigger event, adding the custom instructions to a system message associated with the language model, the system message being a prompt modifier to the language model, in response to receiving a prompt, retrieving the custom instructions from the session specific cache, determining whether the custom instructions are relevant to the prompt, and in response to determining the custom instructions are relevant to the prompt, generating a response to the prompt based on the custom instructions.

First claim

What is claimed is:

1. A system for interacting with a language model using custom instructions, the system comprising at least one processor configured to perform operations comprising:
receiving, through an interface, custom instructions, the custom instructions comprising at least one of personal information or a response type preference;
storing the custom instructions in data storage associated with a user profile;
in response to a trigger event, adding the custom instructions to a system message associated with the language model, the system message being a prompt modifier to the language model;
in response to receiving a prompt, retrieving the custom instructions from the data storage;
determining whether the custom instructions are relevant to the prompt;
in response to determining the custom instructions are relevant to the prompt, generating a response to the prompt based on the custom instructions.

2. The system of claim 1 , wherein determining whether the custom instructions are relevant to the prompt:
extracting components in the custom instructions into structured data;
indexing the structured data with the prompt in a session specific context cache; and
analyzing semantics and context of the custom instructions and the prompt.

3. The system of claim 1 , wherein the interface comprises at least one of a text box, a pop-up window, a user interface, a command line interface, a natural language interface, or an application programming interface.

4. The system of claim 1 , wherein the operations further comprise:
detecting the user profile during a session; and
monitoring prompts to the language model during the session.

5. The system of claim 4 , wherein the operations further comprise:
determining whether the prompts include custom instructions; and
in response to determining the prompts include custom instructions, adding the custom instructions to the system message.

6. The system of claim 5 , wherein determining whether the prompts include custom instructions comprises at least one of:
identifying patterns indicating the presence of custom instructions,
analyzing structure and content of prompts, or
recognizing patterns associated with custom instructions by a fine-tuning process.

7. The system of claim 4 , wherein the operations further comprise:
collecting metadata associated with the prompts;
determining whether the metadata includes custom instructions; and
in response to determining the metadata includes custom instructions, adding the custom instructions to the system message.

8. The system of claim 7 , wherein determining whether the metadata includes custom instructions comprises at least one of utilizing pattern matching algorithms to identify structures associated with custom instructions, analyzing the context of the metadata, or recognizing patterns associated with custom instructions by a fine-tuning process.

9. The system of claim 7 , wherein the metadata comprises at least one of user device location, operating system, or default language.

10. The system of claim 1 , wherein:
the system message is unmodifiable by users; and
adding the custom instructions to the system message comprises concatenating the custom instructions to the system message.

11. A computer-implemented method for interacting with a language model using custom instructions, the method comprising:
receiving, through an interface, custom instructions, the custom instructions comprising at least one of personal information or a response type preference;
storing the custom instructions temporarily within a session specific cache;
in response to a trigger event, adding the custom instructions to a system message associated with the language model, the system message being a prompt modifier to the language model;
in response to receiving a prompt, retrieving the custom instructions from the session specific cache;
determining whether the custom instructions are relevant to the prompt;
in response to determining the custom instructions are relevant to the prompt, generating a response to the prompt based on the custom instructions.

12. The method of claim 11 , wherein determining whether the custom instructions are relevant to the prompt:
extracting components in the custom instructions into structured data;
indexing the structured data with the prompt in a session specific context cache; and
analyzing the semantics and context of the custom instructions and the prompt.

13. The method of claim 11 , further comprising:
detecting the user profile during a session; and
monitoring prompts to the language model during the session.

14. The method of claim 13 , further comprising:
determining whether the prompts include custom instructions; and
in response to determining the prompts include custom instructions, adding the custom instructions to the system message.

15. The method of claim 14 , wherein determining whether the prompts include custom instructions comprises at least one of identifying patterns indicating the presence of custom instructions, analyzing structure and content of prompts, or recognizing patterns associated with custom instructions by a fine-tuning process.

16. The method of claim 13 , further comprising:
collecting metadata associated with the prompts;
determining whether the metadata includes custom instructions; and
in response to determining the metadata includes custom instructions, adding the custom instructions to the system message.

17. The method of claim 16 , wherein determining whether the metadata includes custom instructions comprises at least one of utilizing pattern matching algorithms to identify structures associated with custom instructions, analyzing context of the metadata, or recognizing patterns associated with custom instructions by a fine-tuning process.

18. The method of claim 16 , wherein the metadata comprises at least one of user device location, operating system, or default language.

19. A server deploying a language model, the server comprising:
at least one processor;
a storage location connected to the at least one processor; and
a remote access card connected to the at least one processor and the storage location, wherein the at least one processor is configured to:
receive, through an interface, custom instructions, the custom instructions comprising at least one of personal information or a response type preference;
store the custom instructions temporarily within a session specific cache;
add the custom instructions to a system message associated with the language model, the system message being a prompt modifier to the language model;
in response to receiving a prompt, retrieve the custom instructions from the session specific cache;
determine whether the custom instructions are relevant to the prompt;
in response to being relevant, generate a response to the prompt based on the custom instructions.

20. The server of claim 19 , wherein
the interface comprises at least one of a text box;
the system message is unmodifiable by users; and
adding the custom instructions to the system message comprises concatenating the custom instructions to the system message.

Assignees

• Openai Opco Llc

Inventors

• Deutsch Noah
• Zweig Benjamin
• Lopes Raphael Gontijo
• Soto Bustos Javier
• Taylor Iii Jotham
• Yu Kevin

Classifications

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/08

  Learning methods · CPC title