Question banks & learning assessment
Design practice and assessment around objectives, check coverage, difficulty, and answers, then improve the question bank using actual responses.
Inputs
Learning objectives, topic scope, question types, examples, scoring rules, and response data.
Deliverables
Questions and explanations, scoring rubrics, objective coverage, and item-quality analysis.
Step-by-step workflow
Break the work into verifiable stages, then choose models and supporting tools.
- 01
Build an assessment blueprint
Allocate question types and coverage by objective, distinguishing recall, understanding, and practical application.
- 02
Generate questions and scoring guidance
Request answers, derivations, common mistakes, and scoring criteria alongside each question.
- 03
Verify questions independently
Check numerical questions with code and ensure open-ended rubrics are clear and repeatable.
- 04
Revise using response data
Calculate item difficulty, discrimination, and missingness, then check ambiguity and unfair conditions.
Required model capabilities
Combine models for the actual stages. The directory includes candidates matching one or more of these capabilities.
View matching modelsSupporting tools
Selection & delivery checks
- Verify generated answers independently; an output cannot certify its own correctness.
- Eliminate unintended multiple answers, leaked clues, and ambiguous options.
- Evaluate automated scoring against teacher-scored examples first.
An example request
Write questions, worked answers, and scoring rules from course goals. Verify calculations in Python, analyze responses, and identify items to revise.Find models for this request