LongLaMP: A Benchmark for Personalized Long-form Text Generation
LongLaMP provides a comprehensive benchmark for evaluating personalized long-text generation across 4 diverse tasks:
- Personalized Email Completion
- Personalized Abstract Generation
- Personalized Review Writing
- Personalized Topic Writing
Key features:
- Focuses on generating longer, more complex personalized text (vs. short text)
- Includes user and temporal settings for each task
- Provides datasets, evaluation metrics, and baseline results
- Open source and extensible
Visit http://LongLaMP-benchmark.github.io for more details and to access the benchmark.