# US 2026/0064993 A1 · Modular subsequent generations of dedicated intermediate representations

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0064993 A1
- Document ID: US20260064993A1
- Status: Published application
- Kind: A1
- Application number: 18/818,325
- Filed: 2024-08-28
- Published: 2026-03-05
- Earliest claimed: 2024-08-28
- Assignees: Sap SE
- Applicants: Sap SE
- Inventors: David Kunz
- Web page: https://patentgrep.com/patent/US20260064993A1

## CPC

- G06F 40/56: Natural language generation
- G06F 40/58: Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
- G06F 8/35: model driven
- G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30)
- G06F 40/30: Semantic analysis
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/0475: Generative networks
- G06N 3/08: Learning methods
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

In an example embodiment, a feature tree of all features in a request is generated. This feature tree describes interrelations between features that are dependent upon one another. The request is then broken into multiple, smaller requests. More particularly, each feature is included in its own dedicated request. Based on the dependencies, the smaller requests are then sent to the LLM in an order that preserve the dependency relationships. Intermediate results generated by the LLM in response to the multiple smaller request are then merged into a single intermediate result that can then be passed to the programmatic component for compiling.

## Claims

### Claim 1

A system comprising:
at least one hardware processor; and
a computer-readable medium storing instructions that, when executed by the at least one hardware processor, cause the at least one hardware processor to perform operations comprising:
receiving first natural language text requesting automatic generation of generated text, the first natural language text comprising a plurality of features of the generated text;
for each feature in the plurality of features:
generating a prompt for a corresponding feature;
sending the prompt to a large language model (LLM) receiving, from the LLM, a generated intermediate representation;
merging the generated intermediate representations for the plurality of features into an combined intermediate representation; and
passing the combined intermediate representation to a programmatic component, which validates the combined intermediate representation and converts the combined intermediate representation into a final representation.

### Claim 2 (depends on 1)

The system of claim 1, wherein the final representation is compilable computer code.

### Claim 3 (depends on 1)

The system of claim 1, wherein the operations further comprise:
generating a tree structure based on the first natural language text, the tree structure containing a node corresponding to each feature of the plurality of features, with edges between nodes signifying dependencies between features.

### Claim 4 (depends on 3)

The system of claim 3, wherein the prompts are sent to the LLM in an order determined by the tree structure.

### Claim 5 (depends on 4)

The system of claim 4, wherein the order places prompts corresponding to features that are dependent on other features behind prompts corresponding to the other features.

### Claim 6 (depends on 2)

The system of claim 2, wherein the compilable computer code is in a format that is at least partially proprietary.

### Claim 7 (depends on 6)

The system of claim 6, wherein the compilable computer code is a Core Data Services (CDS) model.

### Claim 8 (depends on 1)

The system of claim 1, wherein the intermediate representation is not compilable.

### Claim 9 (depends on 8)

The system of claim 8, wherein the intermediate representation is a JavaScript Object Notation (JSON) file.

### Claim 10 (depends on 1)

The system of claim 1, wherein the programmatic component automatically corrects one or more errors in the intermediate representation.

### Claim 11 (depends on 1)

The system of claim 1, wherein the plurality of features in the natural language text does not represent all of the features in the natural language text.

### Claim 12 (depends on 11)

The system of claim 11, wherein at least one feature that is not in the plurality of features but that is in the natural language text is excluded from the plurality of features based on a selection of the at least one feature by a user in a graphical user interface.

### Claim 13

A method comprising:
receiving first natural language text requesting automatic generation of text, the first natural language text comprising a plurality of features of the text;
for each feature in the plurality of features:
generating a prompt for a corresponding feature;
sending the prompt to a large language model (LLM) receiving, from the LLM, a generated intermediate representation;
merging the generated intermediate representations for the plurality of features into an combined intermediate representation; and
passing the combined intermediate representation to a programmatic component, which validates the combined intermediate representation and converts the combined intermediate representation into a final representation.

### Claim 14 (depends on 13)

The method of claim 13, wherein the final representation is compilable computer code.

### Claim 15 (depends on 14)

The method of claim 14, further comprising:
generating a tree structure based on the first natural language text, the tree structure containing a node corresponding to each feature of the plurality of features, with edges between nodes signifying dependencies between features.

### Claim 16 (depends on 15)

The method of claim 15, wherein the prompts are sent to the LLM in an order determined by the tree structure.

### Claim 17 (depends on 16)

The method of claim 16, wherein the order places prompts corresponding to features that are dependent on other features behind prompts corresponding to the other features.

### Claim 18

A non-transitory machine-readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:
receiving first natural language text requesting automatic generation of text, the first natural language text comprising a plurality of features of the text;
for each feature in the plurality of features:
generating a prompt for a corresponding feature;
sending the prompt to a large language model (LLM) receiving, from the LLM, a generated intermediate representation;
merging the generated intermediate representations for the plurality of features into an combined intermediate representation; and
passing the combined intermediate representation to a programmatic component, which validates the combined intermediate representation and converts the combined intermediate representation into a final representation.

### Claim 19 (depends on 18)

The non-transitory machine-readable medium of claim 18, wherein the plurality of features in the natural language text does not represent all of the features in the natural language text.

### Claim 20 (depends on 18)

The non-transitory machine-readable medium of claim 18, wherein at least one feature that is not in the plurality of features but that is in the natural language text is excluded from the plurality of features based on a selection of the at least one feature by a user in a graphical user interface.
