Free · no sign-up · no time limit
Ready-made multiple-choice practice quizzes, hand-checked. Take any free quiz online — every answer comes with an explanation, and every quiz page has a full answer key to revise from.
Loading quizzes
Free · no sign-up · no time limit
Ready-made multiple-choice practice quizzes, hand-checked. Take any free quiz online — every answer comes with an explanation, and every quiz page has a full answer key to revise from.
IT
Test your knowledge of transformer architectures, PEFT techniques (LoRA), RLHF, vLLM, and production deployment strategies. Answer each question to see whether you got it right, with an explanation every time.
A free 16-question multiple-choice practice test on IT. Each question has one correct answer and a short explanation, and the full answer key below the player doubles as a study sheet. No account, no time limit — take it as often as you like.
This is a free, 16-question multiple-choice quiz. Answer each question to see whether you got it right, with an explanation for every answer. There is no sign-up and no time limit — take it as many times as you like, and use the answer key above to revise before or after.
Correct answer: AIt freezes the base model weights and injects trainable low-rank rank decomposition matrices into attention layers
LoRA freezes the original pre-trained weights and adds pairs of low-rank matrices to attention projections, drastically reducing the number of trainable parameters and optimizer state memory.
Correct answer: BDPO directly optimizes the policy network using preference data without training a separate reward model
DPO mathematically reformulates the RL objective, directly updating the policy network from preferred/rejected human pairs without training a separate reward model or using complex PPO reinforcement learning.
Correct answer: AFragmentation and overallocation of the Key-Value (KV) cache in GPU VRAM
PagedAttention treats the dynamically growing KV cache like virtual memory pages in an operating system, eliminating memory fragmentation and enabling efficient KV cache sharing across requests.
Correct answer: BA small, fast draft model generates candidate tokens verified in parallel by the target model in a single forward pass
Speculative decoding uses a lightweight draft model to propose multiple tokens quickly; the larger, expensive model verifies or rejects them in parallel in one forward pass, boosting throughput.
Correct answer: AMemory reads and writes between fast GPU SRAM and high-bandwidth memory (HBM)
FlashAttention tiles the computation of softmax and attention matrices, reducing quadratic memory traffic between slow HBM and fast on-chip SRAM to make attention I/O-aware.
Correct answer: BNormalFloat 4 (NF4)
QLoRA introduced NormalFloat 4 (NF4), an information-theoretically optimal quantile quantization scheme designed specifically for normally distributed neural network weights.
Correct answer: BThe sharpness of the probability distribution over the vocabulary prior to sampling
Dividing logits by temperature scales the distribution: lower temperatures (<1.0) peak the distribution around high-probability tokens, while higher temperatures flatten it, increasing diversity.
Correct answer: BPositions outside the trained frequency domain cause severe perplexity degradation unless position interpolation or scaling is applied
RoPE encodes position via complex rotation angles; exceeding trained token lengths causes out-of-distribution rotation frequencies, degrading output quality unless scaled (e.g., via NTK-aware scaling).
Correct answer: BScheduling and evicting individual requests at each token iteration rather than at the full sequence level
Continuous batching inserts new requests and completes finished sequences at every single token generation step, maximizing GPU compute utilization compared to static request batching.
Correct answer: BIt discounts the logits of tokens that have already been generated to discourage loops
Repetition penalty reduces the logit value of previously generated tokens, making the sampling algorithm significantly less likely to select them again.
Correct answer: APrefill processes the prompt tokens in parallel to generate the initial KV cache; decode generates one token at a time autoregressively
The prefill phase is compute-bound, ingesting the full prompt concurrently to populate the KV cache; the subsequent decode phase is memory-bandwidth-bound, emitting one token per step.
Correct answer: BTo establish high-priority behavioral constraints, persona, and guidelines that persist throughout the interaction
System prompts provide foundational framing, safety boundaries, role-playing personas, and instructions that anchor model behavior throughout downstream conversation turns.
Correct answer: BThe loss of previously acquired general capabilities when a model is heavily trained on a narrow domain dataset
Catastrophic forgetting occurs when gradient updates on a specialized task overwrite the general linguistic and reasoning representations learned during large-scale pre-training.
Correct answer: ATime to First Token (TTFT)
Time to First Token (TTFT) measures the latency between dispatching a prompt and receiving the very first streamed token, defining perceived interactive responsiveness.
Correct answer: AA high-performance file format designed for rapid loading and execution on CPUs and GPUs via llama.cpp
GGUF is a extensible binary format created by the llama.cpp project that encapsulates model metadata, tensor data, and quantized weights for cross-platform local execution.
Correct answer: BTo prevent the policy model from drifting too far from the original reference model's generation distribution
A Kullback-Leibler (KL) divergence penalty ensures that the policy being updated with RL does not deviate drastically from the unaligned base model, preventing reward hacking and gibberish outputs.