[ { "id": "code", "kind": "ocr-code", "prompt": "Transcribe the code in this screenshot exactly, then in one sentence state the bug on line 7.", "truth": { "must_contain": [ "merge_windows", "sorted", "gap", "out.append" ], "bug_terms": [ "end", "last[1]", "extend", "max" ] }, "image": "cases/code.png" }, { "id": "chart", "kind": "chart-read", "prompt": "Read this bar chart. Which bar is tallest, which is shortest, and what is the approximate value of each of the five bars?", "truth": { "tallest": "delta", "shortest": "gamma", "values": { "alpha": 12, "beta": 27, "gamma": 9, "delta": 33, "epsilon": 18 }, "tolerance": 4 }, "image": "cases/chart.png" }, { "id": "shapes", "kind": "count-ground", "prompt": "Count the shapes in this image. How many circles, how many squares, and how many triangles? Give the colour of each group. Answer with just the counts and colours.", "truth": { "counts": { "circle": 4, "square": 3, "triangle": 2 }, "colours": { "circle": "red", "square": "blue", "triangle": "green" } }, "image": "cases/shapes.png" }, { "id": "ui", "kind": "ui-read", "prompt": "This is a screenshot of a settings panel. List every setting name and its value, and name the two buttons at the bottom.", "truth": { "must_contain": [ "us-west-2", "3", "Enabled", "16 GiB", "/healthz", "Save", "Cancel" ] }, "image": "cases/ui.png" }, { "id": "text-control", "kind": "text-only", "image": null, "prompt": "In one short paragraph, explain why merging sorted intervals is O(n log n) rather than O(n).", "truth": { "must_contain": [ "sort" ] } } ]