Dataset · lockon
60K Verified Function Calling Examples
lockon/xlam-function-calling-60k
English dataset of 60 thousand verified examples for training LLMs in tool and API use.
What's inside
The dataset contains 60,000 English text records collected using the APIGen automated pipeline. Each example underwent three-step verification: format control, actual function execution, and semantic verification, with 600 samples additionally verified by humans. The data is labeled for text generation, question answering, and reinforcement learning (RL) tasks.
What you can do with it
- SFT fine-tuning of LLMs for reliable external tool and API use (function calling)
- Training AI agents via reinforcement learning on verifiable scenarios
- Evaluating accuracy in generating structured function calls
- Tasks
- Question answering, Text generation, Reinforcement learning
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 41.7K
- Likes
- 1
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("lockon/xlam-function-calling-60k")