LLM Query Augmentation for Word Processor Autoformatting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing large language models (LLMs) integrated with word processing applications do not effectively capture document formatting rules, requiring users to manually adjust the visual presentation of generated text, leading to inefficiencies and increased computational load.
Innovation Solution
A query augmentation engine identifies and incorporates document formatting rules into user queries, enabling LLMs to generate content that adheres to predefined layout and visual constraints, such as paragraph indentation, line breaks, and image placement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If LLMs are integrated with word processing applications for text generation, then content generation capability is improved, but formatting compliance deteriorates
Solution Approach 1:
The system performs preliminary action by extracting and incorporating formatting rules into the prompt before the LLM generates content. The formatting rules are retrieved from the document and integrated into the system prompt that guides the LLM, ensuring that formatting constraints are established beforehand rather than requiring post-generation adjustments.
Solution Approach 2:
The system introduces an intermediary component that acts as a bridge between the LLM and the document formatting requirements. This intermediary retrieves formatting rules from the document and translates them into prompt instructions that the LLM can understand and follow, mediating between the generative capability of the LLM and the structural requirements of the document.
2Speed
If LLMs generate content without formatting rules integration, then generation speed is improved, but manual formatting effort increases
Solution Approach 1:
The system enables self-service by automatically retrieving formatting rules from the document and incorporating them into the generation process without requiring user intervention. The system serves itself by autonomously managing the formatting constraints, eliminating the need for users to manually format the generated content.
Solution Approach 2:
By performing the formatting rule extraction and integration before content generation, the system eliminates the need for subsequent manual formatting work. The preliminary incorporation of formatting constraints into the prompt ensures that the LLM generates content that is immediately ready for use without requiring additional formatting time.
3Manufacturing precision
If formatting rules are incorporated into prompts, then formatting compliance is improved, but computational load increases
Solution Approach 1:
The system applies local quality by selectively extracting and incorporating only the relevant formatting rules that apply to the specific content generation task, rather than including all possible formatting constraints. This targeted approach incorporates formatting precision where needed while minimizing unnecessary computational overhead from irrelevant rules.
4Shape
If users manually adjust formatting of generated text, then visual presentation is improved, but user productivity deteriorates
Solution Approach 1:
The system performs self-service by automatically ensuring formatting compliance through prompt engineering, eliminating the need for users to manually adjust formatting. The LLM generates content that already adheres to the document's formatting rules, so users can directly use the generated content without additional formatting work.
Solution Approach 2:
By incorporating formatting rules into the prompt before generation, the system performs the formatting preparation in advance, so the generated content is ready for immediate use. This preliminary action eliminates the need for users to spend time on manual formatting adjustments, thereby maintaining both visual presentation quality and user productivity.
Data Source
AI summary
Techniques and systems are disclosed that perform content generation with integrated automated formatting using word processing applications that deploy large language models (LLMs). The techniques include receiving, a natural language (NL) query for a synthetic content for a document, identifying formatting rules of the document, and generating an augmented query that includes a representation of at least a portion of the NL query and a representation of the one or more formatting rules of the document. The techniques further include providing the augmented query to an LLM and updating the document with the synthetic content generated by the LLM in response to the augmented query.