Samsung patents system that builds better AI prompts using extra data
Samsung describes a method where a device prepares a specific instruction for a large language model by combining the user's question with a pre-saved template and extra details. The device checks if more information is needed, fetches it from another server, and sends the complete package to the AI server to get a structured answer.
In the filing’s own words
An electronic apparatus, including a memory configured to store a plurality of sample prompts; a communication interface configured to communicate with a server including a large language model; and at least one processor configured to: acquire user input of a user, select a sample prompt corresponding to the user input from among the stored plurality of sample prompts, acquire additional information corresponding to the user input, acquire an input prompt based on the user input, the sample prompt, and the additional information, transmit the input prompt to the server using the communication interface; and based on receiving response information corresponding to the input prompt from the server through the communication interface, provide the response information to the user.
What exactly they claim (claim 1)
1. An electronic apparatus comprising: a communication interface configured to communicate with a first server comprising a language model; and at least one processor configured to: obtain user input acquired from a user, select a sample prompt related to the user input from a stored plurality of sample prompts, wherein the sample prompt includes response structure information indicating a format of information to be included in a response generated based on the sample prompt, determine whether additional information is necessary for providing a response corresponding to the user input, transmit a signal requesting the additional information to a second server using the communication interface based on determining that the additional information is to be used, receive the additional information from the second server using the communication interface, obtain an input prompt by combining the user input acquired from the user with the sample prompt selected by the at least one processor and the additional information, transmit the input prompt to the first server using the communication interface, receive response information from the first server using the communication interface, wherein the response information corresponds to output data that is generated according to the response structure information by providing the input prompt as input to the language model, and provide the response information to the user.
Facts
- Company
- Samsung (SAMSUNG ELECTRONICS CO., LTD.)
- Official title
- Electronic apparatus for generating input prompt for language model, and controlling method thereof
- Patent number
- US12762335B2
- Application number
- 18230393
- Filed
- Aug 4, 2023
- Granted
- Oct 13, 2026
- Inventors
- 2
- Claims
- 12
Classified under
G06F 40/20 Physics & computingMore from Samsung in AI & software
- Samsung patents tool to erase objects from photos using AI Sep 24, 2026
- Samsung patents drawing objects into photos with AI Oct 15, 2026
- Samsung patents AI that creates personalized exercise videos Sep 10, 2026
- Samsung patents AI summaries for paused TV shows based on your viewing habits Sep 24, 2026
- Samsung patents turning chat conversations into image prompts Oct 1, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.