CCAR-F Context and Reliability Practice Question
A financial services firm is using Claude 3.5 Sonnet to analyze quarterly earnings reports that exceed 150,000 tokens. Which TWO strategies are most effective for maintaining high reliability when the model must extract specific data points from the middle of this large context?
⚠ Common exam trap
Many students place queries at the beginning of massive documents or omit structural tags, leading to degraded attention in the middle of long contexts.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Encapsulating documents in XML tags like <document> and <report>
Long-context models like Claude 3.5 Sonnet are highly capable, but performance can still vary based on data placement. Using XML tags helps the model distinguish between different sections of the input, while placing the specific question or 'call to action' at the very end of the prompt ensures the model focuses on the task after processing all context.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Using markdown headers exclusively for sectioning
Why it's wrong here
Markdown headers are helpful for human readability but are less robust than XML tags for Claude's internal processing. XML tags provide clear, unambiguous boundaries that the model is specifically trained to recognize, making them a superior choice for structuring complex, multi-part documents in a way that maximizes data retrieval accuracy.
- ✓
Encapsulating documents in XML tags like <document> and <report>
Why this is correct
XML tags act as clear structural markers that help Claude navigate large context windows. By explicitly labeling sections, you reduce the cognitive load on the model and minimize the 'lost in the middle' phenomenon, where information in the center of a long prompt might otherwise be weighted less heavily.
- ✗
Placing the user query at the beginning of the prompt
Why it's wrong here
Placing the query at the beginning can be counterproductive in very long contexts because the model may lose focus by the time it reaches the end of the 150,000-token document. Standard best practices for Claude suggest that the specific instructions should follow the context to ensure they are the most recent information.
- ✓
Positioning the 'Query' or 'Task' after all reference material
Why this is correct
Putting the task description at the end of the prompt is a proven technique for improving performance in long-context scenarios. This ensures that the instructions are fresh in the model's 'attention' after it has already ingested the bulk of the data, leading to more relevant and reliable extractions.
- ✗
Disabling the system prompt to save token space
Why it's wrong here
The system prompt is essential for setting the persona and operational constraints of the model. Disabling it to save a few hundred tokens in a 150,000-token context is a poor trade-off, as it significantly degrades the model's ability to follow complex instructions and maintain a reliable professional tone.
About these practice questions
Courseiva writes every CCAR-F question from scratch — 271 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCAR-F practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCAR-F exam.