A toy language model has learned -dimensional embeddings. The query word puppy has the vector
and four candidate words sit in the same space:
| Word | Embedding vector |
|---|---|
| dog | |
| kitten | |
| the | |
| wheel |
A retrieval system ranks candidates by cosine similarity to the query and returns the highest-scoring one.
Which word does it return as puppy's nearest neighbour?
Select all that apply.