ECON 100 MACROECONOMICS ACTUAL TEST
PAPER 2026 COMPLETE QUESTIONS AND
CORRECT ANSWERS GRADED A+
▶ Which of the following is NOT a characteristic of PyTorch?
A. Python first
B. Easy to debug
C. Extensive ecosystem
D. Default use of static maps - Answer: D
▶ Which of the following statements is true about the training of large
models?
A. Foundation models cannot be directly deployed.
B. The pre-training of large models requires a large amount of high-quality
data.
C. Unlabeled data used in the pre-training phase of large models must be
used in the fine-tuning phase.
D. Large models have emergent abilities, and therefore they do not need to
be aligned with human values. - Answer: B
▶ Which of the following is not an advanced Al application?
A. Knowledge graph
B. rget Quantum computing
C. SDN
D. Intelligent driving - Answer: C
▶ What are the main schools of thought in Al? (Multiple choice)
A. Symbolism
B. Illusionism
C. Actionism
D. Connectionism - Answer: ACD
▶ What are the layers used in Transformer to capture time dependencies
within a sequence? (Multiple choice)
A. Residual layer
B. Self-attention layer
, C. Location code
D. Feed-forward network layer - Answer: BC
▶ Which of the following hyperparameters may be defined during model
training? (Multiple choice)
A. Batch size
B. Learning rate
C. Optimizer type
D. Loss value - Answer: ABC
▶ Which of the following are common evaluation metrics for classification
tasks? (Multiple choice)
A. Accuracy
B. Precision
C. Recall rate
D. Mean squared error - Answer: ABC
▶ Which of the following statements about data preprocessing are true?
(Multiple choice)
A. Data cleansing is a process of filling in missing values, as well as
detecting and eliminating noise and exceptions.
B. Data dimension reduction aims to simplify data attributes and avoid the
curse of dimensionality.
C. Data standardization aims to reduce noise data and improve model
accuracy by standardizing data.
D. Machine learning tasks generate results in the form of model outputs.
Therefore, model training is more important than data preprocessing. -
Answer: ABC
▶ Which of the following statements about model parameters and
hyperparameters are true? (Multiple choice)
A. Models contain both parameters and hyperparameters.
B. Hyperparameters are automatically learned by models.
C. Hyperparameters are manually set.
D. Hyperparameters can be used to control training. - Answer: ACD
▶ Data augmentation can improve model robustness and avoid overfitting.
Which of the following are data augmentation methods? (Multiple choice)
A. Batch normalization
B. Image contrast change
PAPER 2026 COMPLETE QUESTIONS AND
CORRECT ANSWERS GRADED A+
▶ Which of the following is NOT a characteristic of PyTorch?
A. Python first
B. Easy to debug
C. Extensive ecosystem
D. Default use of static maps - Answer: D
▶ Which of the following statements is true about the training of large
models?
A. Foundation models cannot be directly deployed.
B. The pre-training of large models requires a large amount of high-quality
data.
C. Unlabeled data used in the pre-training phase of large models must be
used in the fine-tuning phase.
D. Large models have emergent abilities, and therefore they do not need to
be aligned with human values. - Answer: B
▶ Which of the following is not an advanced Al application?
A. Knowledge graph
B. rget Quantum computing
C. SDN
D. Intelligent driving - Answer: C
▶ What are the main schools of thought in Al? (Multiple choice)
A. Symbolism
B. Illusionism
C. Actionism
D. Connectionism - Answer: ACD
▶ What are the layers used in Transformer to capture time dependencies
within a sequence? (Multiple choice)
A. Residual layer
B. Self-attention layer
, C. Location code
D. Feed-forward network layer - Answer: BC
▶ Which of the following hyperparameters may be defined during model
training? (Multiple choice)
A. Batch size
B. Learning rate
C. Optimizer type
D. Loss value - Answer: ABC
▶ Which of the following are common evaluation metrics for classification
tasks? (Multiple choice)
A. Accuracy
B. Precision
C. Recall rate
D. Mean squared error - Answer: ABC
▶ Which of the following statements about data preprocessing are true?
(Multiple choice)
A. Data cleansing is a process of filling in missing values, as well as
detecting and eliminating noise and exceptions.
B. Data dimension reduction aims to simplify data attributes and avoid the
curse of dimensionality.
C. Data standardization aims to reduce noise data and improve model
accuracy by standardizing data.
D. Machine learning tasks generate results in the form of model outputs.
Therefore, model training is more important than data preprocessing. -
Answer: ABC
▶ Which of the following statements about model parameters and
hyperparameters are true? (Multiple choice)
A. Models contain both parameters and hyperparameters.
B. Hyperparameters are automatically learned by models.
C. Hyperparameters are manually set.
D. Hyperparameters can be used to control training. - Answer: ACD
▶ Data augmentation can improve model robustness and avoid overfitting.
Which of the following are data augmentation methods? (Multiple choice)
A. Batch normalization
B. Image contrast change
