Providing neural networks with "memories" drives outperformance in large language models
- Jan 18, 2022
- 1 min read
"RETRO" models also enhance explainability as modelers can see which parts of the database have been referenced in a given task
top of page
bottom of page
Comments