A company has a conversational AI assistant that sends requests through Amazon Bedrock to an Anthropic Claude large language model (LLM). Users report that when they ask similar questions multiple tim

Question

Options

• AIncrease the temperature parameter and the top_k parameter.
• BIncrease the temperature parameter. Decrease the top_k parameter.
• CDecrease the temperature parameter. Increase the top_k parameter.
• DDecrease the temperature parameter and the top_k parameter.

• A
  4% (1)
• B
  4% (1)
• C
  14% (4)
• D
  79% (22)

Explanation

Option D is correct because reducing both parameters moves the model toward deterministic, consistent output: a lower temperature compresses the probability distribution so the model strongly favors its top predicted token rather than sampling broadly, and a lower top_k shrinks the candidate token pool at each step, leaving fewer options for the model to randomly choose among.

Why the distractors fail: Option A increases both parameters, amplifying randomness in two directions - the opposite of the goal. Option B increases temperature, which is the primary driver of randomness, making outputs more varied even though the lower top_k partially offsets it. Option C correctly lowers temperature but then increases top_k, reintroducing variance by expanding the pool of tokens the model can sample from.

Memory tip: Think "D for Deterministic" - both parameters Down produces Deterministic responses. Temperature and top_k are both "creativity dials"; turning them both down tightens the model's output toward its most confident, repeatable answer.

No community discussion yet for this question.