Dataset · bowen-upenn
Dataset for LLM personalization and agent memory
bowen-upenn/PersonaMem-v2
10 to 100 thousand English examples for training models to work with hidden user profiles and memory.
What's inside
The set includes 10 to 100 thousand English tabular and text records created for the AI personalization task. The data aims to train language models in extracting hidden user preferences and working with agent memory when solving subjective tasks.
What you can do with it
- Training assistants with long-term user memory
- Customizing personalized response generation to human preferences
- Evaluating the ability of agents to account for implicit user profiles
- Tasks
- Question answering, Text generation
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 17.3K
- Likes
- 40
How to load it
from datasets import load_dataset
ds = load_dataset("bowen-upenn/PersonaMem-v2")