[LLM] Create golden dataset generation tool#481
Merged
Merged
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
|
|
||
| # Use deterministic hashing for consistent assignment | ||
| identifier = user_id or session_id | ||
| hash_value = int(hashlib.md5(f"{experiment_id}:{identifier}".encode()).hexdigest(), 16) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR introduces a Golden Dataset Generation Tool for creating high-quality, reproducible evaluation datasets for LLM workflows. The tool standardizes dataset generation, validation, and versioning, enabling consistent benchmarking of prompts, models, and application behavior across releases.
Changes Made
Added a tool for generating golden datasets from configurable data sources.
Implemented support for customizable dataset generation rules and templates.
Added metadata generation for prompts, expected outputs, model configuration, and evaluation context.
Implemented dataset validation to ensure completeness, consistency, and formatting.
Added deterministic generation options for reproducible datasets.
Introduced dataset versioning and change tracking to support regression testing.
Added CLI/configuration options for dataset generation and export.
Supported export to common formats such as JSON, JSONL, and CSV.
Added structured logging and reporting for generation progress and validation results.
Updated documentation with usage examples and configuration guidance.
Benefits
Standardizes creation of evaluation datasets.
Enables reliable benchmarking across prompts and models.
Supports regression testing with reproducible datasets.
Simplifies maintenance of evaluation data through versioning.
Improves confidence in LLM quality and performance measurements.
Testing
Verified dataset generation from supported input sources.
Confirmed deterministic generation produces identical outputs across repeated runs.
Validated dataset schema, metadata, and export formats.
Tested handling of invalid inputs and generation failures.
Verified versioning and change tracking functionality.
Executed unit and integration tests to ensure reliable operation.
Checklist
Golden dataset generation tool implemented
Configurable generation rules added
Dataset validation implemented
Deterministic generation supported
Dataset versioning and metadata added
Export formats implemented
Documentation updated
Unit and integration tests added and passing..closed #399