Dataset · zjunlp
Predicting ML solution quality before code execution
zjunlp/PredictBeforeExecute
A corpus of ML solutions, agent runs, and analytical artifacts for evaluating algorithm performance without execution.
What's inside
It contains curated ML solution corpora, task resources, agent run results, and analytical artifacts. The dataset belongs to the text modality and includes between 10,000 and 100,000 records. The data was collected to study solution preferences based on data analysis and LLM reasoning.
What you can do with it
- Training or evaluating LLMs for selecting the best ML pipeline without code execution
- Benchmarking ML agents on pre-evaluation tasks of solutions
- Size
- 10K–100K rows
- Downloads this month
- 27.3K
- Likes
- 2
How to load it
from datasets import load_dataset
ds = load_dataset("zjunlp/PredictBeforeExecute")