Skip to content

task_symbolic__radial_code_wheel__code_output_label

1) Identity

  1. Domain: symbolic
  2. Scene id: radial_code_wheel
  3. Task id: task_symbolic__radial_code_wheel__code_output_label
  4. Objective contract: choose the output-label option reached by a three-symbol radial code.

Program Contract

Program: radial_code_wheel.code_to_output_lookup(scene=radial_code_wheel, scope=source_code_and_six_output_options, output=option_label)

Candidate set: the six visible output-label option cards labeled A..F. Operands: the shown three-symbol source code and the three-ring radial code wheel mapping to terminal output labels. Operation: follow the source code through the inner, middle, and outer ring symbols to find the terminal output, then select the option containing that output label. Output binding: answer is the selected option letter. Annotation witnesses: a point_map with inner_ring_symbol, middle_ring_symbol, and outer_ring_symbol centers for the shown code. Query ids: single.

Reasoning Operations

Families: transformation

2) Scene + task contract

  1. Entities/relations:
  2. one three-ring radial code wheel with symbols A|B|C|D;
  3. one visible source code card;
  4. 64 unique terminal output labels on the wheel;
  5. six visual output-label options A..F.
  6. Supported query_id values: single
  7. answer_gt.type: string
  8. Default annotation_gt.type: point_map
  9. Annotation schema: point_map
  10. Alternate annotation forms: none
  11. Annotation witness policy:
  12. minimal object/primitive witnesses: the inner, middle, and outer ring-symbol centers followed by the shown code.
  13. annotation shape choice: point_map, because the task binds three ordered semantic ring roles.
  14. map annotation role names: inner_ring_symbol, middle_ring_symbol, outer_ring_symbol
  15. numeric/readout annotation handling: not applicable.
  16. answer-option annotation policy: the selected option card is not annotated; the answer field already provides the chosen option label.
  17. Overlap/touch policy: terminal labels and option cards are rendered from fixed non-overlapping layout slots.

3) Prompt contract

  1. prompt_bundle_id: symbolic_radial_code_wheel_v1
  2. scene_key: radial_code_wheel
  3. task_key: code_output_label
  4. Optional query-id prompt mapping: none
  5. Required slots:
  6. answer output: json_output_contract_answer_only, answer_hint, json_example_answer_only
  7. answer+annotation mode: json_output_contract, annotation_hint, answer_hint, json_example
  8. JSON example validity rule: examples use a capital option letter answer and a three-role point-map annotation.
  9. Variant counts: 5 scene templates, 5 task templates, 5 output-mode templates.
  10. Output modes: answer_only, answer_and_annotation

4) Determinism + constraints

  1. Seed namespaces used: task dataset, scene variant, prompt variants, and symbolic background/noise namespaces.
  2. Unique-answer policy: all 64 terminal labels are unique; six option labels are unique; exactly one option contains the reached output label.
  3. Reject/resample conditions: invalid code length/symbols, duplicate terminal labels, unsupported option count, or insufficient distractors.
  4. No-auto-relaxation guarantee: invalid constraints raise and retry with the next deterministic retry seed.

5) Tests

  1. Determinism test: tests/test_symbolic_radial_code_wheel_tasks.py
  2. Answer/annotation consistency test: same focused test file validates code-to-terminal output and annotation roles.
  3. Prompt metadata/placeholder test: same focused test file validates prompt bundle keys and required slots.
  4. Constraint-specific tests: same focused test file validates unique terminal/output option construction.