Separate
Score the final answer independently from the learner’s causal explanation.
Correct answer, wrong reason
Compare a learner’s answer and explanation, diagnose the rule behind a lucky answer, then verify a pre-authored synthetic replay response on a different problem.
One product.
One decision.
Visible evidence.
Interactive proof
These exact twenty labeled fraction cases are evaluated by product.py.
Reveal the expected decision generated by the product's deterministic acceptance fixture.
How it works
Score the final answer independently from the learner’s causal explanation.
Map the explanation to a specific misconception, not a generic wrong label.
Verify the fixture’s pre-authored response to a different contrastive problem; mark RESOLVED only when both answer and reasoning are correct.