CCAR-F Context and Reliability Practice Question
A financial-reporting agent must extract figures from quarterly filings and populate a downstream ledger. Auditors require that every extracted number be traceable to its source. Which design most directly satisfies the traceability requirement?
⚠ Common exam trap
The trap here is conflating extraction accuracy with provenance, when a perfectly correct number still fails an audit if its source cannot be demonstrated.
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
✓
Have the model output each figure together with the source document name and the exact quoted sentence containing it.
Traceability means each ledger figure can be traced back to specific source text. Emitting the document name and the exact quoted sentence with every extracted number creates that link at extraction time, so audit review is direct rather than reconstructive. Logging, larger context, and numeric normalization improve debugging, accuracy, or formatting but never bind a value to its origin.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the model's context window so it can ingest the entire filing at once and reduce extraction errors.
Why it's wrong here
Ingesting more text may improve extraction accuracy, yet accuracy and traceability are different properties. A correct number with no recorded origin still fails an audit requirement that every figure map to its source. Window size changes how much the model sees, not whether the system records where each value came from, so it does not satisfy the stated need.
- ✓
Have the model output each figure together with the source document name and the exact quoted sentence containing it.
Why this is correct
Pairing each extracted number with its document name and verbatim source sentence gives auditors a direct path from ledger entry back to the filing text. This makes every figure independently verifiable without re-reading the whole document. Since the requirement is traceability, embedding provenance in the extraction output is the most direct and auditable design choice available.
- ✗
Apply a post-processing regex pass to normalize all numeric formats before writing to the ledger.
Why it's wrong here
Normalization ensures consistent formatting such as decimal separators and currency symbols, which is valuable for ledger integrity. However, it records nothing about origin, so a normalized figure remains untraceable. The auditors need a source link per number, and reformatting addresses presentation only, leaving the provenance requirement entirely unmet in this scenario.
- ✗
Log the raw model request and response payloads to an immutable store for later inspection.
Why it's wrong here
Payload logging supports debugging and forensic review, but it does not attach provenance to individual numbers in the ledger. An auditor examining a specific figure would have to reconstruct which part of a long response produced it. Logging is complementary evidence, not the traceability mechanism the requirement demands, so it leaves the core audit gap open.
About these practice questions
One of 271 original CCAR-F practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. 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.