LaMP
Emerging3papers using it
2024first seen
The 'LaMP' dataset/benchmark is used to evaluate the effectiveness of personalization techniques for large language models by providing structured rewriting trajectories for training reflection modules that align model outputs with user preferences.