UPSC CSE Prelims
Question 83
Which of the following statements with regard to Large Language Models (LLMs) used in machine learning is/are correct?
- LLMs assign probabilities to the next possible words and then pick the one with the highest probability.
- LLMs process data through mathematical optimization to minimise prediction errors.
- LLMs produce unbiased outputs.
Select the answer using the code given below:
View correct answer and explanation
Correct answer: B
Correct Option: B
Explanation: LLMs learn statistical patterns by optimising predictive objectives and generate text via token-probability distributions; they can inherit biases.
Statement-wise Analysis:
Statement 1: Correct - In simplified descriptions, next-token probabilities guide output; actual generation can sample rather than always choose the highest-probability token.
Statement 2: Correct - Training uses mathematical optimisation to reduce losses/prediction error.
Statement 3: Incorrect - Outputs are not automatically unbiased because training data and model behaviour can contain systemic bias.
Option-wise Analysis:
A: Incorrect - Omits supported item(s) 2.
B: Correct - This option includes the supported item(s) and excludes the unsupported ones.
C: Incorrect - Includes unsupported item(s) 3 and omits supported item(s) 1.
D: Incorrect - Includes unsupported item(s) 3.
Review note: Nuance: decoding does not always choose the single highest-probability next token; temperature sampling and other search methods are common.