Free HPE0-V30 sample questions

Correct answer: To prevent the dot products from growing too large, which would push the softmax function into regions with extremely small gradients.

As the dimension of the key vectors (d_k) increases, the variance of the dot product increases, leading to very large values. Feeding these large values into a softmax function pushes the outputs to 1 or 0, resulting in extremely small (vanishing) gradients during backpropagation. Scaling by the square root of d_k normalizes the variance.

Correct answers: The original Transformer paper uses sine and cosine functions of different frequencies to generate the encodings.; They are necessary because the self-attention mechanism is inherently permutation-invariant and has no notion of token order.

The original "Attention Is All You Need" paper proposed using continuous sinusoidal functions (sine and cosine) of varying frequencies to inject positional information, allowing the model to easily learn to attend by relative positions.

Self-attention operations treat the input sequence as a "bag of words." Without positional encoding, the model would not be able to distinguish the order of tokens, which is critical for understanding language syntax and semantics.

Correct answer: Encoder-only architecture utilizing masked language modeling

BERT (Bidirectional Encoder Representations from Transformers) relies on an encoder-only architecture. It uses masked language modeling (MLM) during pre-training, allowing it to look at context from both the left and right simultaneously, unlike autoregressive decoder models (like GPT) that only look at past tokens.

Correct answer: To store high-dimensional embeddings and perform rapid similarity searches.

Vector databases are specialized systems designed to store, manage, and index high-dimensional vector embeddings generated by AI models. Their primary function is to perform extremely fast similarity searches (like approximate nearest neighbor search) to find data semantically related to a user's query.

Correct answer: Passing the text chunks through an embedding model to generate vector representations.

Before text can be stored and searched semantically in a vector database, it must be converted into numerical representations (vectors). This is done by passing the text chunks through an embedding model (like text-embedding-ada-002 or MiniLM). flowchart LR Text[Raw Text] --> Chunk[Chunking] Chunk --> Embed[Embedding Model] Embed --> Vec[Vectors] Vec --> VDB[(Vector DB)]

Correct answer: Inverted File Index (IVF) combined with Product Quantization (PQ), as it significantly compresses the vectors in memory while accelerating search via clustering.

For massive datasets (500M vectors) with strict memory constraints, IVF-PQ is the optimal choice. Product Quantization compresses the vectors, drastically reducing the RAM required. IVF clusters the data so that searches only scan a fraction of the database, achieving the low latency required without the massive memory overhead typical of graph-based indexes like HNSW.