CCAO-F Claude Model Fundamentals Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Increasing the temperature to 1.0

  Why it's wrong here

  Increasing the temperature actually increases the randomness of the model's output, which makes hallucinations more likely to occur. Higher temperature encourages the model to take more creative risks with token selection, which is detrimental when the goal is strict factual adherence to a technical document. (49 words)

• ✗

  Using Claude 3 Haiku for higher precision

  Why it's wrong here

  While Haiku is fast, it generally has lower reasoning capabilities than Sonnet or Opus. For complex technical summarization, a smaller model might be more prone to missing nuances or making errors. Choosing Haiku for precision is incorrect, as Opus and 3.5 Sonnet are the flagship models for accuracy. (51 words)

• ✓

  Providing the documents in the context and requesting citations

  Why this is correct

  By placing the technical specifications in the prompt and asking the model to cite specific passages, you force Claude to ground its response in the provided text. This 'RAG-style' approach ensures the model focuses on the evidence at hand, reducing the likelihood of generating outside or invented information. (52 words)

• ✗

  Setting 'max_tokens' to a very low value

  Why it's wrong here

  Setting a low 'max_tokens' value simply truncates the model's output, potentially cutting off a response mid-sentence. It does nothing to improve the factual accuracy or the quality of the reasoning. In fact, it might prevent the model from providing the necessary context to explain its summary correctly. (50 words)