Computer scienceData scienceNLPLanguage representation

Language Models concepts

LM training strategies and their disadvantages

Report a typo

Match the training strategies for Language models and their disadvantages.

Match the items from left and right columns

Rule-based models

Feature-based models

Transfer learning models

Prompt-based models

The most challenging approach to scale from one domain to another.

Hard to interpret due to complex internal workings and require significant storage space.

Generated text heavily depends on the quality and clarity of prompts, inherit biases and limitations from training data.

Difficult to generalize well for inputs or scenarios significantly different from the training data.

___

Create a free account to access the full topic