Sentence transformers in Prompt Engineering / GenAI - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load a pre-trained sentence transformer model.

Prompt Engineering / GenAI

from sentence_transformers import SentenceTransformer
model = SentenceTransformer('[1]')

Drag options to blanks, or click blank then click option'

Abert-base-uncased

Ball-MiniLM-L6-v2

Cresnet50

Dgpt-3

Attempts:

3 left

💡 Hint

Common Mistakes

Using a model name from a different library like GPT or ResNet.

Using a base BERT model which is not fine-tuned for sentence embeddings.

The correct model name for a popular sentence transformer is all-MiniLM-L6-v2. Other options are either not sentence transformer models or unrelated.

2fill in blank

medium

Complete the code to encode a list of sentences into embeddings.

Prompt Engineering / GenAI

sentences = ['Hello world', 'Machine learning is fun']
embeddings = model.[1](sentences)

Drag options to blanks, or click blank then click option'

Aencode

Btransform

Cpredict

Dfit

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'fit' which is for training models, not encoding sentences.

Using 'predict' which is for classification or regression outputs.

The encode method converts sentences into vector embeddings suitable for similarity tasks.

3fill in blank

hard

Fix the error in the code to compute cosine similarity between two sentence embeddings.

Prompt Engineering / GenAI

from sklearn.metrics.pairwise import [1]
similarity = cosine_similarity([emb1], [emb2])[0][0]

Drag options to blanks, or click blank then click option'

Acosine_similarity

Beuclidean_distances

Cpairwise_distances

Dmanhattan_distances

Attempts:

3 left

💡 Hint

Common Mistakes

Using distance functions which give dissimilarity, not similarity.

Misspelling the function name.

The function cosine_similarity computes the cosine similarity between vectors, which is the correct metric here.

4fill in blank

hard

Fill both blanks to create a dictionary of sentence embeddings for given sentences.

Prompt Engineering / GenAI

sentences = ['AI is amazing', 'I love coding']
embeddings = model.encode(sentences)
embedding_dict = {sentences[[1]]: embeddings[[2]] for i in range(len(sentences))}

Drag options to blanks, or click blank then click option'

Ai

B0

D1

Attempts:

3 left

💡 Hint

Common Mistakes

Using different variables for sentence and embedding indices.

Using fixed indices instead of the loop variable.

We use the loop variable i to index both sentences and embeddings to build the dictionary.

5fill in blank

hard

Fill all three blanks to filter sentences with embeddings having norm greater than 1.0.

Prompt Engineering / GenAI

import numpy as np
filtered = {sent: emb for sent, emb in zip(sentences, embeddings) if np.linalg.[1](emb) [2] [3]

Drag options to blanks, or click blank then click option'

Anorm

B>

C1.0

Dsum

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'sum' instead of 'norm' which sums elements but does not compute vector length.

Using wrong comparison operators.

The np.linalg.norm function computes the length of the embedding vector. We filter embeddings with norm greater than 1.0.