Ex05 and ChatGPT

How can you group constituents into phrases? (i.e., what are the deciding criteria?)

• there are none – constituents are randomly grouped into phrases

• linguistic theory

• by starting letters

• phrases contain constituents that belong together

Which one is the correct c-structure?

• A)

• B)

Which way to specify rules on verb phrases is preferable?

• A)

• B)

We want as many distinct c-structures as there are…

• constituents in the sentence.

• VPs in the sentence.

• tokens in the sentence.

• plausible readings of the sentence.

In f-structures grammars, values of features need to be pushed up to check for…

• spelling.

• word order.

• agreement.

• coherence.

ChatGPT is based on the large language model … .

• Turing NLG

• GPT-3

• GPT-1

• DALL-E 2

Compared to GPT-3, ChatGPT is …

• less conversational.

• more specialized.

• faster.

• more powerful.

Language models try to predict the next token, given …

• the number of tokens it has stored.

• the current weather conditions.

• the already seen sequence.

• the most often used words of the language.

Language models are often separated into the classes …

• transformer vs. attention.

• probabilistic vs. neural.

• attention vs. encoder-decoder.

• neural vs. probabilistic.