Company: Oracle_MCQ_14nov
Difficulty: medium
What settings can be adjusted to minimize randomness in large language model responses for a chatbot?
For a sentiment analysis task with limited resources for fine-tuning, what technique allows a large language model to learn the new task without gradient updates?
A Generative AI engineer is fine-tuning a large language model for a sentiment classification task. It is observed that the gradient computed on the model becomes too small after only a few iterations, and it does not reduce the training loss. What activation functions are used in hidden layers of large language models?
A Translation Lookaside Buffer (TLB) has an access time of 5ns and a hit ratio of 80%. Main memory takes 8 times longer to access than the TLB. On a TLB miss, a main memory access is required for the page table lookup. Assuming there is no page fault, what is the effective access time (in ns)?
The following queue pages are to be arranged in a 3 framed cache: 4, 2, 3, 1, 2, 6, 8, 2, 2, 6. Select the true statement regarding Most Recently Used (MRU) and Least Recently Used (LRU) page replacement.
Code A and Code B share a common critical section. Code A: private p; if(p < 5) wait(); elseif(p == 5) signal(); else enter(critical); Which statement is true regarding Code A?
Given a hash table of length 10, only positive numbers are inserted. Which option is true regarding these two statements? Statement I: h(x) = x * odd natural number mod 10 is a uniform hash function. Statement II: h(x) = x * even natural number mod 10 is a uniform hash function.
An articulation point, also known as a "separating vertex", is one whose removal increases the number of connected components in a graph. How many articulation points are in the graph shown?
A circular linked list can be used to implement:
Which of the following is a valid DFS traversal starting from node 9 of the given graph?