QUIZ 2 MR. JOB

Test

(1)

Parties jouées 45 %Précision 74 Temps moyen 03:55

Créer une copie

À propos de cette activité

HIGH SCHOOL SCIENCE CONFERENCE

Créé par

Laras Sekar Kinanti

Indonesia

Téléchargez la version pour jouer sur papier

Top 10 résultats

#	Jeux	Précision	temps
1	Keira Josephine Wijaya 21 Août 2026	% 100	05:13
2	Anonyme 21 Août 2026	% 100	06:08
3	Anonyme 21 Août 2026	% 93	01:02
4	Anonyme 21 Août 2026	% 93	01:28
5	Muhamad Hafidz Muttaqi_SMAN 10 Bekasi 21 Août 2026	% 93	01:30
6	Anonyme 21 Août 2026	% 93	01:31
7	Anonyme 21 Août 2026	% 93	02:44
8	Anonyme 21 Août 2026	% 93	02:49
9	I Gede Parama Putra 21 Août 2026	% 93	03:02
10	Nabila Audisa 21 Août 2026	% 93	03:08

Voulez-vous apparaître dans le Top 10 de ce jeu? Connectez-vous pour vous identifier.

Créez votre propre jeu gratuite à partir de notre créateur de jeu

Affrontez vos amis pour voir qui obtient le meilleur score dans ce jeu

Créer un défi

UPS 5 Seeing Habits Quiz

Heather Bell

United States

(12)

Test your knowledge about the 5 seeing habits discussed in the text 'UPS 5 seeing habits'.
Practice Quiz - Punnett Square, Variation, and Mutations

Benjamin D Morgan

United States

(14)

Practice Quiz - Punnett Square, Variation, and Mutations
Tissues - Practice Quiz

Benjamin D Morgan

United States

(16)

Tissues - Practice Quiz
Spanish Alphabet Listening

Falkenberg Falkenberg

United States

(88)

Spanish alphabet, listening-based assessment
Photosynthesis vs Cellular Respiration - Practice Quiz

Benjamin D Morgan

United States

(39)

Photosynthesis vs Cellular Respiration - Practice Quiz

		%
	Anonyme	%
		%
		%

Imprimez votre jeu

QUIZ 2 MR. JOB


QUIZ 2 MR. JOBVersion en ligne HIGH SCHOOL SCIENCE CONFERENCE par Laras Sekar Kinanti 1 Why does the Transformer use positional encoding? a To prevent attention scores from being normalized by softmax b To provide sequence-order information that is not inherently supplied by self-attention c To replace token embeddings with a position-dependent vocabulary d To determine which tokens should be classified as unknown 2 What is the main purpose of the attention mechanism in the Transformer architecture? a To dynamically weight contextual information from different input positions when forming representations b To determine the fixed vocabulary size before the model begins training c To encode the original sequence position so that token order is preserved d To convert contextual representations directly into vocabulary indices 3 In the original Transformer architecture, what is the main distinction between the encoder and decoder? a The encoder uses only positional information, while the decoder uses only token embeddings b The encoder performs tokenization, while the decoder converts vocabulary IDs back into text c The encoder processes input representations, while the decoder generates outputs using masked self-attention and encoder-decoder attention d The encoder and decoder use identical attention patterns and differ only in vocabulary size 4 Why do NLP models need a tokenizer before text is given to the model? a Because tokenization guarantees that every sentence has the same number of tokens b Because tokenization removes the need for embeddings and positional information c Because tokenizers perform the same prediction task as language models d Because raw text must be mapped into discrete units that can subsequently be represented numerically 5 What is a major drawback of word-based tokenization when the vocabulary must cover a very large number of words? a It prevents frequent words from receiving individual representations b It can produce a very large vocabulary and struggle with unseen words c It forces every word to be decomposed into individual characters d It removes contextual relationships between neighboring words 6 What is a key advantage of character-based tokenization over word-based tokenization? a It gives every token a richer semantic meaning than a complete word b It guarantees shorter sequences for every input sentence c It eliminates the need for a language model to interpret context d It uses a smaller basic vocabulary and can represent unseen words through characters 7 Why is subword tokenization often considered a compromise between word-based and character-based tokenization? a It keeps many common units intact while allowing uncommon words to be decomposed into smaller units b It assigns one unique token to every complete sentence c It restricts the vocabulary to the most frequent complete words d It converts every word into a fixed number of characters 8 In the tokenizer encoding process, what normally happens after text is split into tokens? a The tokens are reconstructed into a sentence before numerical processing b The tokens are converted into positional encodings without vocabulary lookup c The tokens are mapped to numerical IDs according to the tokenizer's vocabulary d The tokens are discarded after their frequency has been calculated 9 What does decoding mean in a tokenizer? a Learning new vocabulary entries from an already encoded sequence b Applying attention to determine the meaning of each token c Converting raw text into token IDs for the first time d Mapping token or vocabulary IDs back into a human-readable text string 10 In self-attention, why can relationships among tokens be important? a Because each token is assigned the same meaning regardless of its context b Because each token's representation can incorporate information from other tokens in the sequence c Because the vocabulary itself explicitly stores every syntactic relationship d Because self-attention prevents a token from interacting with distant positions 11 What is the general role of softmax in a Transformer? a It converts token IDs into embeddings with fixed dimensionality b It removes positional information before attention is calculated c It transforms scores into normalized values that can serve as attention weights or probabilities d It selects a single vocabulary item before the attention layer is executed 12 What is meant by vocabulary in a tokenizer? a The collection of token units that the tokenizer can map to IDs or representations b The complete set of model parameters learned during training c The set of all sentences encountered during model training d The list of outputs generated by the language model 13 Why is subword tokenization useful for languages with complex words that can be formed from many parts? a It requires the tokenizer to store every possible word form in advance b It ensures that every complex word is always represented by one token c It eliminates the need to learn relationships between token units d It allows complex forms to be represented compositionally rather than requiring every full word to be in the vocabulary 14 What is the relationship between tokenization and input IDs? a Input IDs are created first and determine how the text will be tokenized b Tokenization produces token units, which are then mapped to their corresponding vocabulary IDs c Input IDs are an alternative name for positional encodings d Tokenization is performed only after the model has generated an output 15 What key idea makes the Transformer different from approaches that rely entirely on sequential processing? a It eliminates the need for training data by computing relationships directly from text b It removes numerical representations from the language-processing pipeline c It models relationships between sequence positions through attention, allowing broad interactions without processing every dependency strictly step-by-step d It restricts computation to one token at a time to preserve sequence order

Vous pouvez intégrer le jeu dans un LMS compatible avec LTI 1.1 ou LTI 1.3 comme Canvas, Moodle ou Blackboard. Les scores seront ainsi automatiquement enregistrés dans le carnet de notes de la plateforme.

QUIZ 2 MR. JOB

Test

À propos de cette activité

Créé par

Téléchargez la version pour jouer sur papier

Top Jeux

Test

UPS 5 Seeing Habits Quiz

Test

Practice Quiz - Punnett Square, Variation, and Mutations

Test

Tissues - Practice Quiz

Test

Spanish Alphabet Listening

Test

Photosynthesis vs Cellular Respiration - Practice Quiz

QUIZ 2 MR. JOBVersion en ligne

par Laras Sekar Kinanti

Why does the Transformer use positional encoding?

What is the main purpose of the attention mechanism in the Transformer architecture?

In the original Transformer architecture, what is the main distinction between the encoder and decoder?

Why do NLP models need a tokenizer before text is given to the model?

What is a major drawback of word-based tokenization when the vocabulary must cover a very large number of words?

What is a key advantage of character-based tokenization over word-based tokenization?

Why is subword tokenization often considered a compromise between word-based and character-based tokenization?

In the tokenizer encoding process, what normally happens after text is split into tokens?

What does decoding mean in a tokenizer?

In self-attention, why can relationships among tokens be important?

What is the general role of softmax in a Transformer?

What is meant by vocabulary in a tokenizer?

Why is subword tokenization useful for languages with complex words that can be formed from many parts?

What is the relationship between tokenization and input IDs?

What key idea makes the Transformer different from approaches that rely entirely on sequential processing?