Across
- 1. - Neural network design behind GPT and BERT models.
- 3. - Process of adapting a pre-trained model to a specific domain or task.
- 7. - Converts words or data into dense numeric vectors for similarity tasks.
- 9. - Architecture with two competing networks, often used for image generation.
- 10. - Crafting instructions to guide LLMs toward desired responses.
Down
- 2. - Combines retrieval of external data with generative models for better answers.
- 4. - Field of Al focused on understanding and processing human language.
- 5. - Breaking text into smaller units for model input.
- 6. - Running a trained model to produce predictions or outputs.
- 8. - A type of model with billions of parameters used for text generation.