Quiz 31: BERT Explorer

Test your understanding of BERT's architecture, masked language modeling, next sentence prediction, and how BERT processes and understands language.

📚 Related Learning Tools:

🔧 Interactive Visualization

Check out these interactive tools to learn more before taking the quiz!

Home All Quizzes

1. What does BERT stand for?

Back-End Recurrent Transformer

Bidirectional Encoder Representations from Transformers

Binary Entity Recognition Transformer

Basic Encoding and Recursive Translation

2. What makes BERT different from traditional left-to-right language models?

It reads text bidirectionally to capture full context

It uses convolutional layers instead of attention

It processes words in random order

It only uses tokens from the first sentence

3. What are the two pre-training tasks used by BERT?

Sentiment prediction and summarization

Masked Language Modeling and Next Sentence Prediction

Grammar correction and translation

Word embedding and decoding

4. Which token is used to classify the sentence pair in BERT?

[CLS]

[PAD]

[SEP]

[MASK]

5. What is the purpose of the [MASK] token during training?

To separate two segments

To mark a word that the model should predict

To identify padding locations

To initialize a new sentence

🔒