101. What is the primary advantage of using a pre-trained language model like GPT instead of training a model from scratch?
a) It requires less computational resources.
b) It results in higher accuracy on specific tasks.
c) It reduces the risk of overfitting.
d) It allows for faster training.
102. Which of the following is a disadvantage of using a pre-trained language model like GPT?
a) It may not be suitable for certain tasks.
b) It requires more computational resources than training a model from scratch.
c) It is more difficult to fine-tune than a model trained from scratch.
d) It is less accurate than a model trained from scratch.
103. Which of the following is an example of a task that GPT-3 can perform?
a) Image captioning
b) Speech recognition
c) Language translation
d) Text summarization
104. Which of the following is NOT a characteristic of GPT-3?
a) It is a transformer-based language model.
b) It was trained on a diverse range of text.
c) It can perform multiple tasks without task-specific training.
d) It has a fixed set of parameters that cannot be modified.
105. What is the purpose of the attention mechanism in GPT?
a) To focus on specific words or phrases in the input text
b) To reduce the number of parameters in the model
c) To improve the model’s ability to generate coherent text
d) To speed up the model’s training time
