intentguard
IntentGuard-2
IntentGuard 2 Training Dataset
This dataset contains synthetic training and evaluation examples for IntentGuard, a library for checking whether Python code satisfies natural-language assertions. It was generated with intentguard/ai_research/dataset_generation and follows the same high-level format as the previous IntentGuard dataset, with examples centered on code intent, maintainability, design patterns, reliability, and API behavior.
Each row pairs one natural-language… See the full description on the dataset page: https://huggingface.co/datasets/kdunee/IntentGuard-2.IntentGuard-1
IntentGuard Training Dataset
This dataset is designed to train models for IntentGuard, a novel tool that verifies complex code properties using natural language assertions. It provides a valuable resource for training models to understand the nuanced relationship between natural language descriptions of code intent and actual code behavior. The dataset contains a curated collection of Python code snippets paired with natural language assertions describing desired properties of… See the full description on the dataset page: https://huggingface.co/datasets/kdunee/IntentGuard-1.IntentGuard-2-alpaca-format
IntentGuard 2 Alpaca Format
This dataset is IntentGuard-2 converted to Alpaca instruction format.
Every row uses instruction: "/intentguard".
Source Files
IntentGuard-2/train.jsonl
IntentGuard-2/test.jsonl
Output Files
train.json
test.json
Both output files are JSON arrays. Each item contains instruction, input, and output.
Rebuild
Run python3 scripts/convert_to_alpaca.py from this directory.
IntentGuard-1-alpaca-formatThis is the IntentGuard-1 dataset converted to alpaca instruction format.