Page-Context Actions Using Field Placeholders for LLM Privacy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud computing environments lack the ability to implement page context-aware actions efficiently, particularly in multi-tenant database systems, where user interactions and data privacy are crucial, and existing technologies do not effectively leverage large language models (LLMs) to provide context-aware actions without compromising user data privacy.
Innovation Solution
A system and method that utilizes a planner and a large language model (LLM) to generate context-aware actions by determining web page context and user input, while maintaining data privacy by not transmitting actual field values, and executing actions based on field placeholders.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If actual field values are transmitted to LLM for context-aware actions, then action accuracy is improved, but user data privacy is compromised
Solution Approach 1:
The patent extracts only the essential contextual information (field placeholders, page structure, action types) from the web page while deliberately excluding actual field values containing sensitive user data. This selective extraction allows the LLM to understand page context and generate accurate actions without exposure to private information.
Solution Approach 2:
The patent introduces an intermediary processing layer that transforms actual field values into anonymized placeholders before transmitting to the LLM. This intermediary mechanism preserves the structural and contextual information needed for accurate action generation while acting as a barrier that prevents sensitive data from reaching the external model.
2Measurement precision
If comprehensive web page context is provided to LLM, then context-aware action quality is improved, but transmission data volume increases
Solution Approach 1:
The system extracts only the essential contextual elements required for action generation, such as field placeholders, page structure information, and action type definitions, while omitting redundant or unnecessary data. This selective extraction maintains action quality while minimizing transmission volume.
Solution Approach 2:
The patent transmits a partial representation of the web page context that contains just enough information for the LLM to generate accurate actions. Rather than transmitting the complete page content, only the critical structural and contextual elements are included, achieving sufficient action quality with reduced data volume.
3Object-affected harmful factors
If field placeholders are used instead of actual values, then data privacy is protected, but system complexity increases
Solution Approach 1:
The patent extracts and transmits field placeholder information that represents the structure and semantics of data fields without containing actual values. This approach adds minimal processing complexity while effectively protecting user privacy through systematic value anonymization.
Solution Approach 2:
The field placeholder mechanism serves multiple functions simultaneously: it protects user privacy by masking actual values, preserves contextual information for accurate action generation, and provides a standardized interface for LLM processing. This multi-functionality justifies the modest increase in system complexity.
Data Source
AI summary
Disclosed are some implementations of systems, apparatus, methods and computer program products for implementing page context aware actions. Input submitted via an input device is obtained. A context is determined based upon a web age rendered via a display device, where the context includes one or more field placeholders. The input is transmitted to a large language model (LLM). The determined context is transmitted to the LLM, wherein values of the one or more field placeholders are not transmitted to the LLM. A plan including at least one of the field placeholders is received from the LLM, where the plan includes one or more actions, at least one action being with respect to one of the field placeholders. One or more field values are substituted into the field placeholders of the plan. The plan is executed by executing the one or more actions of the plan, at least one action of the one or more actions being with respect to a field value of the one or more field values.