Q18Fundamentals of GenAI
A company intends to add generative AI capabilities to its application by integrating a large language model (LLM). The LLM responses must be as deterministic and stable as possible. Which solution fulfills these requirements?
← → navigate · a answer
Community votes
Discussion · 2
A 3
The temperature parameter in large language models (LLMs) controls the randomness of the output:
Lower temperature (e.g., 0) → More deterministic and consistent responses
Higher temperature (e.g., 1) → More creative, diverse, and less predictable responses
Setting the temperature to 0 forces the model to always choose the most likely next word, making outputs as deterministic and stable as possible — exactly what the company requires.
A 1
Lower temp ---> Deterministic and stable responses.
Higher temp ---> Increases randomness and variability.