ADD: Upload VLMs strategies prompts and LLM _As_A_Judge prompts

LLM-as-a-Judge /answer_examples_en.json

@@ -0,0 +1,1786 @@
+{
+  "bb_m_01.png": {
+    "correct_examples": [
+      {
+        "left_answer": "smooth coastline",
+        "right_answer": "ribbed coastline"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "\"regular\" shapes",
+        "right_answer": "irregular shapes"
+      }
+    ]
+  },
+  "bb_m_02.png": {
+    "correct_examples": [
+      {
+        "left_answer": "elongated objects",
+        "right_answer": "non elongated objects"
+      },
+      {
+        "left_answer": "longer",
+        "right_answer": "shorter"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Rectangular objects",
+        "right_answer": "Round objects"
+      }
+    ]
+  },
+  "bb_m_03.png": {
+    "correct_examples": [
+      {
+        "left_answer": "object (island or lake) fits into the frame fully",
+        "right_answer": "no islands or lakes present fully in the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "there is a small island in each picture",
+        "right_answer": "most of the pictures are occupied by greenery"
+      }
+    ]
+  },
+  "bb_m_04.png": {
+    "correct_examples": [
+      {
+        "left_answer": "cars present",
+        "right_answer": "immovable objects"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "parking lots",
+        "right_answer": "land plot"
+      }
+    ]
+  },
+  "bb_m_08.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Orientation of cars along one straight line",
+        "right_answer": "Cars are not directed along one straight line"
+      },
+      {
+        "left_answer": "cars parallel",
+        "right_answer": "cars not parallel"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "enlarged photograph, car parking, vehicles",
+        "right_answer": "reduced photograph, car parking, vehicles"
+      },
+      {
+        "left_answer": "car parking",
+        "right_answer": "road"
+      }
+    ]
+  },
+  "bb_m_09.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Solid curved lines ",
+        "right_answer": "Branched lines"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Solid curved lines ",
+        "right_answer": "Branched lines"
+      }
+    ]
+  },
+  "bb_m_10.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Artificially marked land plots",
+        "right_answer": "Natural irregularly shaped landscapes"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "field",
+        "right_answer": "forest"
+      }
+    ]
+  },
+  "bb_m_100.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Line",
+        "right_answer": "Cross"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "top-down view",
+        "right_answer": "angled view"
+      }
+    ]
+  },
+  "bb_m_101.png": {
+    "correct_examples": [
+      {
+        "left_answer": "texture stripes are located approximately vertically",
+        "right_answer": "texture stripes are located at different angles"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Solid rock",
+        "right_answer": "Sand"
+      }
+    ]
+  },
+  "bb_m_102.png": {
+    "correct_examples": [
+      {
+        "left_answer": "scene objects form vertical lines",
+        "right_answer": "scene objects form horizontal lines"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": " Field",
+        "right_answer": "Forested area"
+      }
+    ]
+  },
+  "bb_m_103.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Orientation diagonally upward to the right",
+        "right_answer": "Orientation diagonally downward to the left"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "arbitrary baseball field",
+        "right_answer": "even baseball field"
+      }
+    ]
+  },
+  "bb_m_104.png": {
+    "correct_examples": [
+      {
+        "left_answer": "There is an elongated curve",
+        "right_answer": "There is an elongated straight line"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Houses predominately",
+        "right_answer": "Nature"
+      }
+    ]
+  },
+  "bb_m_105.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Pieces of forest among open spaces",
+        "right_answer": "Open spaces among forest"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "more natural colors",
+        "right_answer": "less natural colors"
+      }
+    ]
+  },
+  "bb_m_106.png": {
+    "correct_examples": [
+      {
+        "left_answer": "land is surrounded by water",
+        "right_answer": "water is surrounded by land"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "wild beach",
+        "right_answer": "not wild beach"
+      }
+    ]
+  },
+  "bb_m_107.png": {
+    "correct_examples": [
+      {
+        "left_answer": "more even shore",
+        "right_answer": "uneven shore"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Even coastlines, the overall shade of the photographs is green",
+        "right_answer": "Curved/winding coastlines, the overall shade of the photographs is blue"
+      }
+    ]
+  },
+  "bb_m_108.png": {
+    "correct_examples": [
+      {
+        "left_answer": "complex coastline",
+        "right_answer": "even coastline"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "smooth separation boundaries",
+        "right_answer": "Straight separation boundaries"
+      }
+    ]
+  },
+  "bb_m_109.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Coast of an urbanized agglomeration",
+        "right_answer": "Coast of a suburb or uninhabited territories"
+      },
+      {
+        "left_answer": "bridges",
+        "right_answer": "no bridges"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "river",
+        "right_answer": "lake"
+      }
+    ]
+  },
+  "bb_m_11.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Buildings",
+        "right_answer": "Wild area"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "No trees",
+        "right_answer": "Trees"
+      }
+    ]
+  },
+  "bb_m_110.png": {
+    "correct_examples": [
+      {
+        "left_answer": "city",
+        "right_answer": "steppe"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Circular square ",
+        "right_answer": "Ring road"
+      }
+    ]
+  },
+  "bb_m_111.png": {
+    "correct_examples": [
+      {
+        "left_answer": "special tracks",
+        "right_answer": "real road interchanges"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Country highway",
+        "right_answer": "City highway"
+      }
+    ]
+  },
+  "bb_m_112.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Pools are located on the northern side",
+        "right_answer": "Pools are located on the southern side"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "green is represented in a narrow dynamic color range",
+        "right_answer": "green is represented in a wide dynamic color range"
+      }
+    ]
+  },
+  "bb_m_113.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Object with a quadrilateral in the center",
+        "right_answer": "Object with a circle in the center"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Pronounced inner cutout in the roofs of the central buildings",
+        "right_answer": "Solid roofs on the central buildings"
+      }
+    ]
+  },
+  "bb_m_114.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Naturally created road",
+        "right_answer": "Road artificially created by humans"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Natural landscapes",
+        "right_answer": "Artificial landscapes"
+      }
+    ]
+  },
+  "bb_m_115.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Power transmission towers",
+        "right_answer": "Wind power plants"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "no shadows",
+        "right_answer": "shadows present"
+      }
+    ]
+  },
+  "bb_m_116.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Animals in the distance",
+        "right_answer": "Animals nearby"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Tree shadows are cast onto the pasture, sheep shadows are not distinguishable",
+        "right_answer": "Sheep shadows are distinguishable and clearly visible"
+      }
+    ]
+  },
+  "bb_m_117.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Buildings",
+        "right_answer": "Nature"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "pond",
+        "right_answer": "ground"
+      }
+    ]
+  },
+  "bb_m_118.png": {
+    "correct_examples": [
+      {
+        "left_answer": "grouped in one place ",
+        "right_answer": "not grouped in one place "
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "sheep",
+        "right_answer": "goats"
+      }
+    ]
+  },
+  "bb_m_12.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Water",
+        "right_answer": "Forests"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Nature",
+        "right_answer": "fields"
+      }
+    ]
+  },
+  "bb_m_120.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Same-type vehicles",
+        "right_answer": "Vehicles of different types"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "several cars",
+        "right_answer": "trucks"
+      }
+    ]
+  },
+  "bb_m_121.png": {
+    "correct_examples": [
+      {
+        "left_answer": "presence of clouds",
+        "right_answer": "absence of clouds"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Urbanised area, lots of clouds",
+        "right_answer": "Rural area, few clouds"
+      }
+    ]
+  },
+  "bb_m_122.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Camels",
+        "right_answer": "Non-camels"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "desert",
+        "right_answer": "savanna"
+      }
+    ]
+  },
+  "bb_m_124.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Transport",
+        "right_answer": "Nature"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Planes",
+        "right_answer": "a sheep"
+      }
+    ]
+  },
+  "bb_m_125.png": {
+    "correct_examples": [
+      {
+        "left_answer": "all individuals are identical",
+        "right_answer": "there are distinct individuals"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "animals together",
+        "right_answer": "animals running apart"
+      }
+    ]
+  },
+  "bb_m_13.png": {
+    "correct_examples": [
+      {
+        "left_answer": "areas without objects",
+        "right_answer": "areas with objects"
+      },
+      {
+        "left_answer": "empty",
+        "right_answer": "objects present"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "No transport",
+        "right_answer": "There is transport"
+      }
+    ]
+  },
+  "bb_m_14.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Quadrilaterals",
+        "right_answer": "Circles"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "quadrilaterals",
+        "right_answer": "triangles"
+      }
+    ]
+  },
+  "bb_m_15.png": {
+    "correct_examples": [
+      {
+        "left_answer": "there are line segments",
+        "right_answer": "there are no line segments"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "more water",
+        "right_answer": "less water"
+      }
+    ]
+  },
+  "bb_m_18.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Manmade",
+        "right_answer": "Nature"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "regular shapes",
+        "right_answer": "irregular shapes"
+      }
+    ]
+  },
+  "bb_m_19.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Cloudy",
+        "right_answer": "Clear"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Big city, more clouds",
+        "right_answer": "Suburb, less clouds"
+      }
+    ]
+  },
+  "bb_m_20.png": {
+    "correct_examples": [
+      {
+        "left_answer": "there are trees",
+        "right_answer": "there are no trees"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "horizontal lines",
+        "right_answer": "vertical lines"
+      }
+    ]
+  },
+  "bb_m_22.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Vertical straight lines",
+        "right_answer": "Horizontal straight lines"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Highway",
+        "right_answer": "main road"
+      }
+    ]
+  },
+  "bb_m_23.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Non-white objects",
+        "right_answer": "White images"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "fewer lines",
+        "right_answer": "more lines"
+      }
+    ]
+  },
+  "bb_m_24.png": {
+    "correct_examples": [
+      {
+        "left_answer": "vortex twisting to the right",
+        "right_answer": "vortex twisting to the left"
+      },
+      {
+        "left_answer": "Northern Hemisphere",
+        "right_answer": "Southern Hemisphere"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "vortex",
+        "right_answer": "spiral"
+      }
+    ]
+  },
+  "bb_m_27.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Cold region",
+        "right_answer": "warm region"
+      },
+      {
+        "left_answer": "winter",
+        "right_answer": "summer"
+      },
+      {
+        "left_answer": "snow present",
+        "right_answer": "snow absent"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Mountain landscape",
+        "right_answer": "Plain landscape"
+      }
+    ]
+  },
+  "bb_m_28.png": {
+    "correct_examples": [
+      {
+        "left_answer": "there are leaves",
+        "right_answer": "there are no leaves"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Summer",
+        "right_answer": "winter"
+      }
+    ]
+  },
+  "bb_m_29.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Interchange for four roads",
+        "right_answer": "Interchange for three roads"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "intechange",
+        "right_answer": "road crossing"
+      }
+    ]
+  },
+  "bb_m_126.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Images from videogames",
+        "right_answer": "Real images"
+      },
+      {
+        "left_answer": "3d graphics",
+        "right_answer": "remote sensing imagery"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "One \"type\" of surface touches the border of the image (whether water or land)",
+        "right_answer": "Two \"types\" of surface touch the border of the image (land and forest, water and land)"
+      }
+    ]
+  },
+  "bb_m_30.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Circles",
+        "right_answer": "Squares"
+      },
+      {
+        "left_answer": "circular objects",
+        "right_answer": "Rectangular objects"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Solid curved lines ",
+        "right_answer": "Branched lines"
+      }
+    ]
+  },
+  "bb_m_32.png": {
+    "correct_examples": [
+      {
+        "left_answer": "left-hand traffic",
+        "right_answer": "right-hand traffic"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "agricultural fields",
+        "right_answer": "natural objects"
+      }
+    ]
+  },
+  "bb_m_33.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Baseball field",
+        "right_answer": "Tennis court"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Road without turns",
+        "right_answer": "Intersection"
+      }
+    ]
+  },
+  "bb_m_36.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Freshly plowed land, little vegetation",
+        "right_answer": "Field has sprouted, lots of vegetation"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "chaos",
+        "right_answer": "orderliness"
+      }
+    ]
+  },
+  "bb_m_37.png": {
+    "correct_examples": [
+      {
+        "left_answer": "no snow",
+        "right_answer": "there is snow"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "airplanes",
+        "right_answer": "private houses"
+      }
+    ]
+  },
+  "bb_m_38.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Waves",
+        "right_answer": "Calm"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "yellow markings",
+        "right_answer": "white markings"
+      }
+    ]
+  },
+  "bb_m_41.png": {
+    "correct_examples": [
+      {
+        "left_answer": "pool surrounded by buildings on all sides",
+        "right_answer": "pool not surrounded by buildings on all sides"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "More than 1 boat",
+        "right_answer": "1 boat"
+      }
+    ]
+  },
+  "bb_m_45.png": {
+    "correct_examples": [
+      {
+        "left_answer": "sports fields",
+        "right_answer": "fields"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Most likely large scale",
+        "right_answer": "Most likely small scale"
+      }
+    ]
+  },
+  "bb_m_46.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Top view shot",
+        "right_answer": "Angled shot"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Low-rise houses",
+        "right_answer": "High-rise buildings"
+      }
+    ]
+  },
+  "bb_m_49.png": {
+    "correct_examples": [
+      {
+        "left_answer": "bridge",
+        "right_answer": "not a bridge"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "2 standout objects",
+        "right_answer": "3 contrastively standout objects"
+      }
+    ]
+  },
+  "bb_m_51.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Moving vessels ",
+        "right_answer": "Vessels at parking"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Interchange with circles",
+        "right_answer": "Interchange without circles"
+      }
+    ]
+  },
+  "bb_m_53.png": {
+    "correct_examples": [
+      {
+        "left_answer": "central road markings are yellow",
+        "right_answer": "central road markings are white"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Cars are located parallel to each other",
+        "right_answer": "Cars are located diagonally (perpendicularly)"
+      }
+    ]
+  },
+  "bb_m_54.png": {
+    "correct_examples": [
+      {
+        "left_answer": "\"zebra\" crossings are even rectangles",
+        "right_answer": "zebra crossings\" are represented as beveled (not at 90 degrees)"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Pedestrian crossings",
+        "right_answer": "Pedestrian crossings"
+      }
+    ]
+  },
+  "bb_m_56.png": {
+    "correct_examples": [
+      {
+        "left_answer": "roadside is marked with a \"dashed\" line",
+        "right_answer": "roadside is marked with a solid line (or not marked)"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "single-lane traffic",
+        "right_answer": "two-lane traffic"
+      }
+    ]
+  },
+  "bb_m_57.png": {
+    "correct_examples": [
+      {
+        "left_answer": "dashed",
+        "right_answer": "solid"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Single line without branches",
+        "right_answer": "All lines have branches"
+      }
+    ]
+  },
+  "bb_m_58.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Incomplete figures",
+        "right_answer": "Complete figures"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "I don't know",
+        "right_answer": "I don't know"
+      }
+    ]
+  },
+  "bb_m_59.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Wild forest",
+        "right_answer": "Forest in the city"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Larger scale. Fewer buildings.",
+        "right_answer": "Smaller scale. More buildings."
+      }
+    ]
+  },
+  "bb_m_60.png": {
+    "correct_examples": [
+      {
+        "left_answer": "pool",
+        "right_answer": "lake, closed reservoir"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "boundaries of water bodies have predominantly angular contours",
+        "right_answer": "boundaries of water bodies have predominantly rounded contours"
+      }
+    ]
+  },
+  "bb_m_61.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Straight road",
+        "right_answer": "Circular road"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Roads are visible in detail",
+        "right_answer": "Roads are visible small"
+      }
+    ]
+  },
+  "bb_m_62.png": {
+    "correct_examples": [
+      {
+        "left_answer": "road",
+        "right_answer": "rails"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Several boats",
+        "right_answer": "One boat"
+      }
+    ]
+  },
+  "bb_m_63.png": {
+    "correct_examples": [
+      {
+        "left_answer": "vertical airplane shots",
+        "right_answer": "horizontal airplane shots"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "looks like computer graphics",
+        "right_answer": "looks like real photographs"
+      }
+    ]
+  },
+  "bb_m_65.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Roofs of houses of the same color",
+        "right_answer": "Roofs of houses of different colors"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "ruler-like layout",
+        "right_answer": "chaotically built buildings"
+      }
+    ]
+  },
+  "bb_m_66.png": {
+    "correct_examples": [
+      {
+        "left_answer": "chaotic form of vehicle parking",
+        "right_answer": "all vehicles are lined up"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "different means of transport",
+        "right_answer": "means of transport are grouped "
+      }
+    ]
+  },
+  "bb_m_67.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Straight roads",
+        "right_answer": "Curved roads"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "circle in the city",
+        "right_answer": "circle in the field"
+      }
+    ]
+  },
+  "bb_m_68.png": {
+    "correct_examples": [
+      {
+        "left_answer": "many lanes",
+        "right_answer": "2 lanes"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Airplanes on the runway",
+        "right_answer": "Residential houses"
+      }
+    ]
+  },
+  "bb_m_69.png": {
+    "correct_examples": [
+      {
+        "left_answer": "three colors",
+        "right_answer": "two colors"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "animals of the same color",
+        "right_answer": "animals of different colors"
+      }
+    ]
+  },
+  "bb_m_70.png": {
+    "correct_examples": [
+      {
+        "left_answer": "presence of four circular objects",
+        "right_answer": "objects intersect each other with straight lines  "
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Animals in the distance",
+        "right_answer": "Animals nearby"
+      }
+    ]
+  },
+  "bb_m_71.png": {
+    "correct_examples": [
+      {
+        "left_answer": "intersections",
+        "right_answer": "circles"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Symmetrically",
+        "right_answer": "Asymmetrically"
+      }
+    ]
+  },
+  "bb_m_73.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Railway",
+        "right_answer": "Road"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Industrial district of the city",
+        "right_answer": "City bridges "
+      }
+    ]
+  },
+  "bb_m_74.png": {
+    "correct_examples": [
+      {
+        "left_answer": "The overpass goes from the bottom border to the top border of the image",
+        "right_answer": "The overpass goes from the left border to the right border of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "More than one vehicle",
+        "right_answer": "No more than one vehicle"
+      }
+    ]
+  },
+  "bb_m_75.png": {
+    "correct_examples": [
+      {
+        "left_answer": "The river crosses the frame relatively straight",
+        "right_answer": "The river makes at least 1 turn greater than 30 degrees"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Road intersections",
+        "right_answer": "Roundabout intersections"
+      }
+    ]
+  },
+  "bb_m_76.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Grid-like, square-nested structure ",
+        "right_answer": "Structure with radial distribution"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Area with reservoirs ",
+        "right_answer": "Area without reservoirs"
+      }
+    ]
+  },
+  "bb_m_79.png": {
+    "correct_examples": [
+      {
+        "left_answer": "island",
+        "right_answer": "peninsula"
+      },
+      {
+        "left_answer": "coastline is closed within frame",
+        "right_answer": "coasstline crosses the frame"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Smooth coastline",
+        "right_answer": "Broken coastline"
+      }
+    ]
+  },
+  "bb_m_80.png": {
+    "correct_examples": [
+      {
+        "left_answer": "natural irregularities ",
+        "right_answer": "artificial irregularities"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "uniform landscape, especially in color",
+        "right_answer": "diverse landscape in color"
+      }
+    ]
+  },
+  "bb_m_83.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Satellite image",
+        "right_answer": "Schematic map"
+      },
+      {
+        "left_answer": "a real photo",
+        "right_answer": "not a real photo"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "mountains",
+        "right_answer": "quarry"
+      }
+    ]
+  },
+  "bb_m_87.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Airports",
+        "right_answer": "Railway stations"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "there are areas with rounded corners",
+        "right_answer": "there are areas with sharp corners"
+      }
+    ]
+  },
+  "bb_m_89.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Symmetrical interchange system",
+        "right_answer": "Asymmetrical interchange system"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "4 \"circles\"",
+        "right_answer": "2 \"circles\""
+      }
+    ]
+  },
+  "bb_m_91.png": {
+    "correct_examples": [
+      {
+        "left_answer": "roads intersect the circle",
+        "right_answer": "roads do not intersect the circle"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "highway",
+        "right_answer": "roundabout, circle"
+      }
+    ]
+  },
+  "bb_m_92.png": {
+    "correct_examples": [
+      {
+        "left_answer": "The horizontal road passes under the vertical one.",
+        "right_answer": "The vertical road passes under the horizontal one"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "circular interchanges",
+        "right_answer": "bridges"
+      }
+    ]
+  },
+  "bb_m_94.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Without ships",
+        "right_answer": "With ships"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "base near water",
+        "right_answer": "parking on water for yachts"
+      }
+    ]
+  },
+  "bb_m_95.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Cars on the left side of the buildings",
+        "right_answer": "Cars on the right side of the buildings"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "there are clouds",
+        "right_answer": "there are no clouds"
+      }
+    ]
+  },
+  "bb_m_96.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Trees in the upper part ",
+        "right_answer": "trees in the lower part "
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Completely forest territories",
+        "right_answer": "There are residential territories"
+      }
+    ]
+  },
+  "bb_m_97.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Bright red objects are located predominantly in the left part/half of the image.",
+        "right_answer": "Bright red objects are located predominantly in the right part/half of the image."
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "multi-lane road",
+        "right_answer": "two-lane road"
+      }
+    ]
+  },
+  "bb_m_98.png": {
+    "correct_examples": [
+      {
+        "left_answer": "The shore stretches from the upper left corner to the lower right",
+        "right_answer": "The shore stretches from the lower left corner to the upper right"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Ordinary road intersections",
+        "right_answer": "Roundabout intersections "
+      }
+    ]
+  },
+  "bb_m_99.png": {
+    "correct_examples": [
+      {
+        "left_answer": "there are no objects inside the circle",
+        "right_answer": "there are objects inside the circle"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Road at nadir  ",
+        "right_answer": "Road at an angle"
+      }
+    ]
+  },
+  "bb_r_38.png": {
+    "correct_examples": [
+      {
+        "left_answer": "The reservoir is enclosed",
+        "right_answer": "The reservoir is not bounded by the image borders on at least one side "
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Dark water",
+        "right_answer": "Light water"
+      }
+    ]
+  },
+  "bb_r_48.png": {
+    "correct_examples": [
+      {
+        "left_answer": "multi-story buildings over 10 floors",
+        "right_answer": "low-rise houses"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "rounded boundaries",
+        "right_answer": "straight boundaries\n"
+      }
+    ]
+  },
+  "bb_r_99.png": {
+    "correct_examples": [
+      {
+        "left_answer": "the center of the circle is empty ",
+        "right_answer": "non-empty center of the circle"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Roundabout traffic",
+        "right_answer": "Roundabout traffic"
+      }
+    ]
+  },
+  "bb_s_05.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Several people",
+        "right_answer": "one person"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "All animals in the image are of the same color",
+        "right_answer": "Each image contains animals of at least 2 colors"
+      }
+    ]
+  },
+  "bb_s_08.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Cars on the left",
+        "right_answer": "Cars on the right"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Organized herd",
+        "right_answer": "Chaotic herd"
+      }
+    ]
+  },
+  "bb_s_11.png": {
+    "correct_examples": [
+      {
+        "left_answer": "There are several cars in the image",
+        "right_answer": "There is only one car in the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "there is parking for cars",
+        "right_answer": "there is no parking for cars"
+      }
+    ]
+  },
+  "bb_s_12.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Cars are visually positioned vertically",
+        "right_answer": "Cars are visually positioned horizontally"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "land areas with reservoirs ",
+        "right_answer": "land areas without reservoirs "
+      }
+    ]
+  },
+  "bb_s_14.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Transport in the left part of the image",
+        "right_answer": "Transport in the right part of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "boat",
+        "right_answer": "boat"
+      }
+    ]
+  },
+  "bb_s_15.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Transport in the upper part of the image",
+        "right_answer": "Transport in the lower part of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "ship",
+        "right_answer": "ship"
+      }
+    ]
+  },
+  "bb_s_17.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Many ships",
+        "right_answer": "Few ships"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Pier",
+        "right_answer": "Boat afloat"
+      }
+    ]
+  },
+  "bb_s_18.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Ships are vertically oriented",
+        "right_answer": "Ships are horizontally oriented"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Airports",
+        "right_answer": "Railway stations"
+      }
+    ]
+  },
+  "bb_s_19.png": {
+    "correct_examples": [
+      {
+        "left_answer": "presence of a tree in the photograph",
+        "right_answer": "absence of trees"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "airplanes",
+        "right_answer": "watercraft"
+      }
+    ]
+  },
+  "bb_s_20.png": {
+    "correct_examples": [
+      {
+        "left_answer": "trees in the upper part of the image",
+        "right_answer": "trees in the lower part of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Mountains without snow",
+        "right_answer": "Snow-covered mountains"
+      }
+    ]
+  },
+  "bb_s_22.png": {
+    "correct_examples": [
+      {
+        "left_answer": "trees",
+        "right_answer": "bushes"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "group of buildings  ",
+        "right_answer": "single buildings"
+      }
+    ]
+  },
+  "bb_s_23.png": {
+    "correct_examples": [
+      {
+        "left_answer": "at least two отдельно standing trees",
+        "right_answer": "one отдельно standing tree"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "cloverleaf interchanges ",
+        "right_answer": "merge-type interchanges"
+      }
+    ]
+  },
+  "bb_s_25.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Presence of at least one lake",
+        "right_answer": "Absence of reservoirs"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Green mountains",
+        "right_answer": "White mountains"
+      }
+    ]
+  },
+  "bb_s_28.png": {
+    "correct_examples": [
+      {
+        "left_answer": "an irregularly shaped reservoir occupies most of the photo",
+        "right_answer": "an irregularly shaped reservoir occupies a small part of the photo"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Wide road",
+        "right_answer": "Narrow road"
+      }
+    ]
+  },
+  "bb_s_29.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Several reservoirs",
+        "right_answer": "One reservoir"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "airplanes",
+        "right_answer": "private houses"
+      }
+    ]
+  },
+  "bb_s_30.png": {
+    "correct_examples": [
+      {
+        "left_answer": "horizontal shot",
+        "right_answer": "vertical shot"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Natural reservoirs",
+        "right_answer": "Artificial reservoirs"
+      }
+    ]
+  },
+  "bb_s_34.png": {
+    "correct_examples": [
+      {
+        "left_answer": "land transport",
+        "right_answer": "water transport"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "intersection",
+        "right_answer": "circle, roundabout"
+      }
+    ]
+  },
+  "bb_s_35.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Absence of trees",
+        "right_answer": "Separately standing tree"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "similar rectangular objects",
+        "right_answer": "absence of groups of objects"
+      }
+    ]
+  },
+  "bb_s_38.png": {
+    "correct_examples": [
+      {
+        "left_answer": "ponds and lakes fully in frame",
+        "right_answer": "open borders water bodies"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "trees",
+        "right_answer": "water"
+      }
+    ]
+  },
+  "bb_s_39.png": {
+    "correct_examples": [
+      {
+        "left_answer": "reservoirs with boats",
+        "right_answer": "reservoirs without boats"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Urban reservoirs",
+        "right_answer": "Suburban reservoirs"
+      }
+    ]
+  },
+  "bb_s_41.png": {
+    "correct_examples": [
+      {
+        "left_answer": "buildings",
+        "right_answer": "presence of trees"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "House",
+        "right_answer": "Garage"
+      }
+    ]
+  },
+  "bb_s_42.png": {
+    "correct_examples": [
+      {
+        "left_answer": "photographs of runway sections",
+        "right_answer": "no runways"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Parallelism ",
+        "right_answer": "Perpendicularity "
+      }
+    ]
+  },
+  "bb_s_43.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Airplanes",
+        "right_answer": "Water transport"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Satellite map images",
+        "right_answer": "Schematic map images"
+      }
+    ]
+  },
+  "bb_s_44.png": {
+    "correct_examples": [
+      {
+        "left_answer": "2 objects",
+        "right_answer": "3 objects"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "lower quality",
+        "right_answer": "higher quality"
+      }
+    ]
+  },
+  "bb_s_49.png": {
+    "correct_examples": [
+      {
+        "left_answer": "buildings",
+        "right_answer": "absence of buildings"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Sports fields",
+        "right_answer": "\"Agricultural\" fields"
+      }
+    ]
+  },
+  "bb_s_53.png": {
+    "correct_examples": [
+      {
+        "left_answer": "private sector",
+        "right_answer": "lonely houses"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "America",
+        "right_answer": "Rich district"
+      }
+    ]
+  },
+  "bb_s_55.png": {
+    "correct_examples": [
+      {
+        "left_answer": "airplanes",
+        "right_answer": "I don't know"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Road for aircraft",
+        "right_answer": "Road for ground vehicles"
+      }
+    ]
+  },
+  "bb_s_57.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Airplane in the upper part of the image",
+        "right_answer": "Airplane in the lower part of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "airplane",
+        "right_answer": "airplane"
+      }
+    ]
+  },
+  "bb_s_58.png": {
+    "correct_examples": [
+      {
+        "left_answer": "several airplanes in the photo",
+        "right_answer": "one airplane in the photo"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "Civil airplanes",
+        "right_answer": "Civil airport"
+      }
+    ]
+  },
+  "bb_s_59.png": {
+    "correct_examples": [
+      {
+        "left_answer": "Airplanes occupy a large area of the image",
+        "right_answer": "Airplanes occupy a small area of the image"
+      }
+    ],
+    "incorrect_examples": [
+      {
+        "left_answer": "curved shore",
+        "right_answer": "straight shore"
+      }
+    ]
+  }
+}

LLM-as-a-Judge /judge_system_prompt_en.txt

@@ -0,0 +1,23 @@
+You evaluate the user's answer by comparing it with reference answers for a given task.
+
+You see:
+• Reference answers
+• Examples of correct and incorrect answers
+• User answer
+
+Each answer contains features of the right and left class. Identify the target property that distinguishes the right class from the left. This may describe semantics, form, presence, quantity, or size of objects.
+
+The user answer may contain:
+• Features of both right and left classes, or
+• An explicit target property separating these classes.
+
+If two features are indicated, formulate the target property based on them.
+
+Answer correctness:
+• correct — if the target property matches the reference or correct examples, accounting for generalization, paraphrasing, synonyms, word order changes, or simplification.
+• incorrect — if the target property is missing or wrong.
+
+Do not accept answers similar to or matching incorrect examples. Focus on meaning, not exact wording.
+Ignore minor differences in style, grammar, or phrasing.
+
+Respond with only one word: correct or incorrect.

LLM-as-a-Judge /reference_answers_en.json

@@ -0,0 +1,498 @@
+{
+  "bb_m_01.png": {
+    "left_answer": "shore as a smooth line",
+    "right_answer": "shore as a jagged line"
+  },
+  "bb_m_02.png": {
+    "left_answer": "elongated object",
+    "right_answer": "non-elongated object"
+  },
+  "bb_m_03.png": {
+    "left_answer": "object boundary fits into frame fully",
+    "right_answer": "object boundaries doesn't fit into frame"
+  },
+  "bb_m_04.png": {
+    "left_answer": "vehicles present",
+    "right_answer": "there are no vehicles"
+  },
+  "bb_m_08.png": {
+    "left_answer": "all cars are parked parallel",
+    "right_answer": "cars parked at different angles"
+  },
+  "bb_m_09.png": {
+    "left_answer": "continuous curved lines",
+    "right_answer": "branched lines"
+  },
+  "bb_m_10.png": {
+    "left_answer": "farm field",
+    "right_answer": "nature"
+  },
+  "bb_m_11.png": {
+    "left_answer": "city",
+    "right_answer": "nature"
+  },
+  "bb_m_12.png": {
+    "left_answer": "water present",
+    "right_answer": "water absent"
+  },
+  "bb_m_13.png": {
+    "left_answer": "absence of objects",
+    "right_answer": "presence of objects"
+  },
+  "bb_m_14.png": {
+    "left_answer": "squares",
+    "right_answer": "circles"
+  },
+  "bb_m_15.png": {
+    "left_answer": "presence of bridges",
+    "right_answer": "absence of bridges"
+  },
+  "bb_m_18.png": {
+    "left_answer": "artificial reservoirs",
+    "right_answer": "natural reservoirs"
+  },
+  "bb_m_19.png": {
+    "left_answer": "presence of haze, clouds",
+    "right_answer": "absence of haze, clouds"
+  },
+  "bb_m_20.png": {
+    "left_answer": "forests",
+    "right_answer": "fields"
+  },
+  "bb_m_22.png": {
+    "left_answer": "vertical roads",
+    "right_answer": "horizontal roads"
+  },
+  "bb_m_23.png": {
+    "left_answer": "colored objects",
+    "right_answer": "there are no colored objects"
+  },
+  "bb_m_24.png": {
+    "left_answer": "twists clockwise from the center",
+    "right_answer": "twists counterclockwise from the center"
+  },
+  "bb_m_27.png": {
+    "left_answer": "presence of snow",
+    "right_answer": "absence of snow"
+  },
+  "bb_m_28.png": {
+    "left_answer": "green trees",
+    "right_answer": "bare trees"
+  },
+  "bb_m_29.png": {
+    "left_answer": "four roads",
+    "right_answer": "three roads"
+  },
+  "bb_m_30.png": {
+    "left_answer": "round objects",
+    "right_answer": "square objects"
+  },
+  "bb_m_32.png": {
+    "left_answer": "left-hand traffic",
+    "right_answer": "right-hand traffic"
+  },
+  "bb_m_33.png": {
+    "left_answer": "baseball field",
+    "right_answer": "tennis court"
+  },
+  "bb_m_36.png": {
+    "left_answer": "plowed field, without greenery",
+    "right_answer": "seeded field"
+  },
+  "bb_m_37.png": {
+    "left_answer": "mountains without snow",
+    "right_answer": "mountains with snow"
+  },
+  "bb_m_38.png": {
+    "left_answer": "wavy sea with foam",
+    "right_answer": "calm sea without foam"
+  },
+  "bb_m_41.png": {
+    "left_answer": "pool surrounded by houses",
+    "right_answer": "pool not surrounded by houses"
+  },
+  "bb_m_45.png": {
+    "left_answer": "sports fields",
+    "right_answer": "agricultural fields"
+  },
+  "bb_m_46.png": {
+    "left_answer": "top-down view",
+    "right_answer": "45 degree view"
+  },
+  "bb_m_48.png": {
+    "left_answer": "multi-story buildings",
+    "right_answer": "low-rise buildings"
+  },
+  "bb_m_49.png": {
+    "left_answer": "bridge",
+    "right_answer": "road"
+  },
+  "bb_m_51.png": {
+    "left_answer": "boats are moving",
+    "right_answer": "boats are stationary"
+  },
+  "bb_m_53.png": {
+    "left_answer": "A yellow marking in the center of the road",
+    "right_answer": "A white marking only in the center of the road"
+  },
+  "bb_m_54.png": {
+    "left_answer": "straight pedestrian crossing",
+    "right_answer": "diagonal crossing"
+  },
+  "bb_m_56.png": {
+    "left_answer": "dashed lines along the edges of the road",
+    "right_answer": "solid lines along the edges of the road"
+  },
+  "bb_m_57.png": {
+    "left_answer": "dashed lines in the middle of the road",
+    "right_answer": "solid lines in the middle of the road"
+  },
+  "bb_m_58.png": {
+    "left_answer": "houses are not enclosed",
+    "right_answer": "enclosed houses"
+  },
+  "bb_m_59.png": {
+    "left_answer": "forest",
+    "right_answer": "park"
+  },
+  "bb_m_60.png": {
+    "left_answer": "pool",
+    "right_answer": "lake"
+  },
+  "bb_m_61.png": {
+    "left_answer": "straight highway",
+    "right_answer": "interchange"
+  },
+  "bb_m_62.png": {
+    "left_answer": "highway",
+    "right_answer": "railway"
+  },
+  "bb_m_63.png": {
+    "left_answer": "the nose of the airplane is directed upward",
+    "right_answer": "the nose of the airplane is directed to the right"
+  },
+  "bb_m_65.png": {
+    "left_answer": "all roofs are of the same color",
+    "right_answer": "there are roofs of different colors"
+  },
+  "bb_m_66.png": {
+    "left_answer": "objects are directed in different directions",
+    "right_answer": "objects are directed in one direction"
+  },
+  "bb_m_67.png": {
+    "left_answer": "straight road",
+    "right_answer": "turning road"
+  },
+  "bb_m_68.png": {
+    "left_answer": "multi-lane road",
+    "right_answer": "single-lane road"
+  },
+  "bb_m_69.png": {
+    "left_answer": "boundary of three colors",
+    "right_answer": "boundary of two colors"
+  },
+  "bb_m_70.png": {
+    "left_answer": "cloverleaf interchange",
+    "right_answer": "diamond interchange"
+  },
+  "bb_m_71.png": {
+    "left_answer": "cross intersection",
+    "right_answer": "roundabout"
+  },
+  "bb_m_73.png": {
+    "left_answer": "railway bridge over cars",
+    "right_answer": "car bridge over the railway"
+  },
+  "bb_m_74.png": {
+    "left_answer": "the bridge over the road is vertical",
+    "right_answer": "the bridge over the road is horizontal"
+  },
+  "bb_m_75.png": {
+    "left_answer": "the river is straight",
+    "right_answer": "the river turns"
+  },
+  "bb_m_76.png": {
+    "left_answer": "city block development",
+    "right_answer": "radial development"
+  },
+  "bb_m_79.png": {
+    "left_answer": "island",
+    "right_answer": "peninsula"
+  },
+  "bb_m_80.png": {
+    "left_answer": "mountains",
+    "right_answer": "quarries"
+  },
+  "bb_m_83.png": {
+    "left_answer": "real image",
+    "right_answer": "map image"
+  },
+  "bb_m_87.png": {
+    "left_answer": "airport",
+    "right_answer": "railway station"
+  },
+  "bb_m_89.png": {
+    "left_answer": "symmetrical interchange",
+    "right_answer": "non-symmetrical interchange"
+  },
+  "bb_m_91.png": {
+    "left_answer": "roundabout with a bridge",
+    "right_answer": "roundabout without a bridge"
+  },
+  "bb_m_92.png": {
+    "left_answer": "the interchange bridge is vertical",
+    "right_answer": "the interchange bridge is horizontal"
+  },
+  "bb_m_94.png": {
+    "left_answer": "port without ships",
+    "right_answer": "port with ships"
+  },
+  "bb_m_95.png": {
+    "left_answer": "car to the left of the house",
+    "right_answer": "car to the right of the house"
+  },
+  "bb_m_96.png": {
+    "left_answer": "forest at the top",
+    "right_answer": "forest at the bottom"
+  },
+  "bb_m_97.png": {
+    "left_answer": "red object on the left",
+    "right_answer": "red object on the right"
+  },
+  "bb_m_98.png": {
+    "left_answer": "diagonal of the coastline from the lower right corner to the upper left",
+    "right_answer": "diagonal of the coastline from the lower left to the upper right"
+  },
+  "bb_m_99.png": {
+    "left_answer": "there is an object inside the roundabout",
+    "right_answer": "the inside of the roundabout is empty"
+  },
+  "bb_m_100.png": {
+    "left_answer": "a road",
+    "right_answer": "an intersection"
+  },
+  "bb_m_101.png": {
+    "left_answer": "vertical patterns on the sand",
+    "right_answer": "diagonal patterns on the sand"
+  },
+  "bb_m_102.png": {
+    "left_answer": "vertical lines",
+    "right_answer": "horizontal lines"
+  },
+  "bb_m_103.png": {
+    "left_answer": "a baseball field in the lower left corner",
+    "right_answer": "a baseball field in the upper right corner"
+  },
+  "bb_m_104.png": {
+    "left_answer": "a turning road",
+    "right_answer": "a straight road"
+  },
+  "bb_m_105.png": {
+    "left_answer": "the forest boundary is convex",
+    "right_answer": "the forest boundary is concave"
+  },
+  "bb_m_106.png": {
+    "left_answer": "the coast boundary is convex",
+    "right_answer": "the coast boundary is concave"
+  },
+  "bb_m_107.png": {
+    "left_answer": "sandy beach",
+    "right_answer": "rocky shore"
+  },
+  "bb_m_108.png": {
+    "left_answer": "the coast is rounded",
+    "right_answer": "the coast is straight"
+  },
+  "bb_m_109.png": {
+    "left_answer": "urbanized river bank",
+    "right_answer": "river bank with greenery"
+  },
+  "bb_m_110.png": {
+    "left_answer": "intersection/interchange in the city",
+    "right_answer": "intersection/interchange outside the city"
+  },
+  "bb_m_111.png": {
+    "left_answer": "racing track",
+    "right_answer": "interchange"
+  },
+  "bb_m_112.png": {
+    "left_answer": "pool above the house",
+    "right_answer": "pool below the house"
+  },
+  "bb_m_113.png": {
+    "left_answer": "rectangle on the roof",
+    "right_answer": "circle on the roof"
+  },
+  "bb_m_114.png": {
+    "left_answer": "rivers",
+    "right_answer": "roads"
+  },
+  "bb_m_115.png": {
+    "left_answer": "power transmission line tower",
+    "right_answer": "wind power plant"
+  },
+  "bb_m_116.png": {
+    "left_answer": "sheep are small",
+    "right_answer": "sheep are large"
+  },
+  "bb_m_117.png": {
+    "left_answer": "city",
+    "right_answer": "animals"
+  },
+  "bb_m_118.png": {
+    "left_answer": "sheep in one cluster",
+    "right_answer": "sheep in several clusters"
+  },
+  "bb_m_120.png": {
+    "left_answer": "only passenger cars or only trucks",
+    "right_answer": "both passenger cars and trucks"
+  },
+  "bb_m_121.png": {
+    "left_answer": "presence of haze, clouds",
+    "right_answer": "absence of haze, clouds"
+  },
+  "bb_m_122.png": {
+    "left_answer": "camels",
+    "right_answer": "zebras"
+  },
+  "bb_m_124.png": {
+    "left_answer": "inanimate objects",
+    "right_answer": "animals"
+  },
+  "bb_m_125.png": {
+    "left_answer": "animals of one color",
+    "right_answer": "animals of different colors"
+  },
+  "bb_m_126.png": {
+    "left_answer": "3d render",
+    "right_answer": "real imaging"
+  },
+  "bb_s_05.png": {
+    "left_answer": "many people",
+    "right_answer": "one person"
+  },
+  "bb_s_08.png": {
+    "left_answer": "cars on the left",
+    "right_answer": "cars on the right"
+  },
+  "bb_s_11.png": {
+    "left_answer": "many cars",
+    "right_answer": "one car"
+  },
+  "bb_s_12.png": {
+    "left_answer": "cars vertically",
+    "right_answer": "cars horizontally"
+  },
+  "bb_s_14.png": {
+    "left_answer": "ships on the left",
+    "right_answer": "ships on the right"
+  },
+  "bb_s_15.png": {
+    "left_answer": "ships at the top",
+    "right_answer": "ships at the bottom"
+  },
+  "bb_s_17.png": {
+    "left_answer": "many ships",
+    "right_answer": "one ship"
+  },
+  "bb_s_18.png": {
+    "left_answer": "ships vertically",
+    "right_answer": "ships horizontally"
+  },
+  "bb_s_19.png": {
+    "left_answer": "there are trees",
+    "right_answer": "there are no trees"
+  },
+  "bb_s_20.png": {
+    "left_answer": "trees at the top",
+    "right_answer": "trees at the bottom"
+  },
+  "bb_s_22.png": {
+    "left_answer": "large trees",
+    "right_answer": "small trees"
+  },
+  "bb_s_23.png": {
+    "left_answer": "many trees",
+    "right_answer": "one tree"
+  },
+  "bb_s_25.png": {
+    "left_answer": "there are lakes",
+    "right_answer": "there are no lakes"
+  },
+  "bb_s_28.png": {
+    "left_answer": "large lakes",
+    "right_answer": "small lakes"
+  },
+  "bb_s_29.png": {
+    "left_answer": "many lakes",
+    "right_answer": "one lake"
+  },
+  "bb_s_30.png": {
+    "left_answer": "vertical lakes",
+    "right_answer": "horizontal lakes"
+  },
+  "bb_s_34.png": {
+    "left_answer": "cars",
+    "right_answer": "ships"
+  },
+  "bb_s_35.png": {
+    "left_answer": "cars",
+    "right_answer": "trees"
+  },
+  "bb_r_48.png": {
+    "left_answer": "high-rise buildings",
+    "right_answer": "low-rise buildings"
+  },
+  "bb_r_38.png": {
+    "left_answer": "lakes",
+    "right_answer": "body of water not closed within the frame"
+  },
+  "bb_r_99.png": {
+    "left_answer": "inside of the roundabout is empty",
+    "right_answer": "there is something inside the roundabout"
+  },
+  "bb_s_39.png": {
+    "left_answer": "ships",
+    "right_answer": "bodies of water without ships"
+  },
+  "bb_s_41.png": {
+    "left_answer": "houses",
+    "right_answer": "trees"
+  },
+  "bb_s_42.png": {
+    "left_answer": "airplanes",
+    "right_answer": "houses"
+  },
+  "bb_s_43.png": {
+    "left_answer": "airplanes",
+    "right_answer": "ships"
+  },
+  "bb_s_44.png": {
+    "left_answer": "two identical objects",
+    "right_answer": "three identical objects"
+  },
+  "bb_s_49.png": {
+    "left_answer": "houses present",
+    "right_answer": "houses absent"
+  },
+  "bb_s_53.png": {
+    "left_answer": "many houses",
+    "right_answer": "one house"
+  },
+  "bb_s_55.png": {
+    "left_answer": "airplanes present",
+    "right_answer": "airplanes absent"
+  },
+  "bb_s_57.png": {
+    "left_answer": "airplanes at the top",
+    "right_answer": "airplanes at the bottom"
+  },
+  "bb_s_58.png": {
+    "left_answer": "several airplanes",
+    "right_answer": "one airplane"
+  },
+  "bb_s_59.png": {
+    "left_answer": "large airplanes",
+    "right_answer": "small airplanes"
+  }
+}

LLM-as-a-Judge /task_type_judge_prompts_en.json

@@ -0,0 +1,23 @@
+{
+  "size": {
+    "prompt": "'You are given answers to a problem of the ''size'' type. Your task is to check whether each answer is correct.\n\nA correct answer must explicitly mention one of the following differences between the objects in the left and right sets of images:\n\n1. The objects in one set are closer, and the objects in the other set are farther.\n\n2. The objects in one set are bigger, and the objects in the other set are smaller.\n\n3. Each set has a name (e.g., ''barns'' on the left, ''estates'' on the right), and these names reflect a real‑world size difference.\nExample: Barns are typically smaller than estates, so stating that difference would be correct logic.\n\nImportant:\n\nIf an answer contains the phrase ''I don't know'' (or any similar expression of uncertainty), it is incorrect.\n\nOutput your judgment as: Correct or Incorrect for each answer, with a brief justification.'"
+  },
+  "presence": {
+    "prompt": "'In this task, the reference target property is based on the presence of an object of a certain class in one set of images and the absence of that same object in the other set. The property is NOT based on the shape of objects, their color, size, or spatial arrangement. However, the target property can be rephrased in semantic terms while preserving meaning. For example, the property ''images in the left class contain clouds, images in the right class do not contain clouds'' can be rephrased semantically as ''the left class depicts cloudy weather, and the right class depicts clear weather,'' because clear weather means the absence of clouds. Analyze the correct answers to understand which semantic rephrasings have already been accepted as valid. Analyze the incorrect answers to understand which semantic formulations lead to errors.\n\nIf the answer to be checked is formulated in terms of presence or absence of an object, evaluate it according to the standard rules. If the answer is formulated in semantic terms, first check whether such semantics has appeared in examples of correct or incorrect answers. If it has, follow those examples. If the semantics is new and has not appeared before, analyze on your own: do the named properties always imply the presence or absence of the target object? If yes — mark the answer as correct, if no — mark it as incorrect.\n\nAdditional evaluation rules: \n\nFirst, if the answer refers to ''image'' in the singular (e.g., ''this image contains X'') rather than ''images'' as a set, you still accept it as valid, as long as the distinction between the two sets is preserved in meaning. Answer is still valid if it includes additional infromation like cause and effect relationship, but incorrect if it includes dditional distinguishing features.\n\nSecond, if the answer describes a characteristic of an object being ''more'' or ''less'' of something (e.g., ''more clouds,'' ''less water'') rather than simply the presence or absence of the target object, you accept it only if no additional distinguishing features are provided beyond that quantitative difference. If the answer includes other differences beyond the more/less comparison, you reject it as incorrect.'"
+  },
+  "number": {
+    "prompt": "'You are given answers to a problem of the ''number'' type. Your task is to check whether each answer is correct.\n\nA correct answer must explicitly mention one of the following differences between the objects in the left and right sets of images:\n\n1. Every image in one set contains more (or fewer) specific objects than every image in the other set.\nExample: Left‑set images each have more apples, right‑set images each have fewer apples.\n\n2. Every image in one set contains a specific fixed number of objects, and every image in the other set contains a different specific fixed number of objects.\nExample: Left set always has 3 cars, right set always has 7 cars.\n\nImportant rules:\n\nIf an answer contains ''I don't know'' or any similar expression of uncertainty, it is incorrect.'"
+  },
+  "shape": {
+    "prompt": "'In this task, the reference target property is based on the shape of objects in the image, regardless of their semantics. However, the target property can be rephrased in terms of object semantics while preserving meaning. For example, the property ''the boundary between land and water is closed or not closed'' can be rephrased semantically as ''island or peninsula.'' Analyze the correct answers to understand which semantic rephrasings have already been accepted as valid. Also analyze the incorrect answers to understand which semantic formulations are mistaken.\n\nIf the answer to be checked is formulated in terms of shape, evaluate it according to the standard rules. If the answer is formulated in terms of semantics, first check whether such semantics has appeared in examples of correct or incorrect answers. If it has, follow those examples. If the semantics is new and has not appeared before, analyze on your own: do objects with such semantics always have the required shape difference? If yes — the answer is correct, if no — it is incorrect.'"
+  },
+  "spatial": {
+    "prompt": "'In this task, the reference target property is based exclusively on the position of objects in the image (absolute location), or on the orientation of objects relative to the image frame (orientation), or on their mutual arrangement relative to each other. The property is NOT based on shape, color, size, quantity, or semantics. Semantic rephrasings are allowed as long as they preserve the meaning of spatial arrangement. For example, ''left-hand traffic'' means that cars move on the left side of the road, and ''right-hand traffic'' means they move on the right side.\n\nGeneral rule: if an answer does not refer to the location or orientation of objects at all, it is incorrect. If the answer refers to a different type of spatial relation (for example, the user writes about the position of an object relative to the image frame, while the correct answer is about the mutual arrangement of two objects), it is also incorrect. Answer is still valid if it includes additional infromation like cause and effect relationship of this spatial relation, but incorrect if it includes additional objects as a distinguishing feature.\n\nFor tasks about absolute location (position of an object or objects within the image): the answer must clearly indicate that the difference between the classes lies in the location. It is not necessary to specify exact parts of the image for each class, but the answer must state what the difference is (for example, ''on different sides of the picture'' is acceptable, whereas simply ''in different places'' is not, because it does not specify what the difference is). The specific object can be referred to generically (''object''). If the object is named specifically, it must be named correctly. Rephrasings are allowed (e.g., ''forest'' instead of ''trees'').\n\nFor tasks about orientation: the answer must clearly indicate that the difference between the classes lies in the orientation. It is not necessary to specify the exact orientation for each class, but the answer must state what the difference is (for example, ''oriented horizontally and vertically'' is acceptable, whereas simply ''oriented differently'' is not, because it does not specify what the difference is). The specific object can be referred to generically (''object''). If the object is named specifically, it must be named correctly. Rephrasings are allowed (e.g., ''forest'' instead of ''trees'').\n\nFor tasks about mutual arrangement of objects: the answer must clearly indicate that the difference lies in the relative arrangement, not in absolute positions. If both objects (or classes of objects) are correctly identified, it is sufficient to state that there is a difference in their mutual arrangement (e.g., ''the car and the tree are on opposite sides of each other''). If specific mutual arrangements are given (e.g., ''the car is to the left of the tree''), they must be correct. It is acceptable to name only one object, but then the answer must specifically state that the difference is in its position relative to the other object. If no object is named, the answer is not counted as correct.\n\nWhen checking, first determine the type of the reference answer (absolute location, orientation, or mutual arrangement), then evaluate the user's answer according to these rules. Analyze the correct answers to understand which formulations have already been accepted as valid. Analyze the incorrect answers to understand common mistakes. \n'"
+  },
+  "same": {
+    "prompt": "'For this task the target property should be shared visual characteristic that unifies objects in one class and sets apart in the other class\nCorrect answer must describe a visual attribute that is uniform across the images in one set and non-uniform in the other collage. \nThe key distinction should rely on intra-set consistency (for example all objects facing the same direction, all objects having the same color or shape) versus variability within the opposing set.\n\nA target property is correct only if it describes a visual attribute that is:\n-present in every image of one class (left or right), and\n- absent from every image of the opposite class, or present in a clearly different consistent form (e.g., all same color vs. multiple colors).\n\nIf an answer says “left images have X, right images have Y” but X is not true for all left images (e.g., “orange roofs” when only some left images have orange roofs), the answer is incorrect — even if Y is correct.\n\nMinor variations in shade, orientation, or size are allowed, but categorical features (color, shape, presence/absence, number, arrangement) must be uniform within each class as stated.'"
+  },
+  "semantic": {
+    "prompt": "'Correct answer identifies a difference in object type, purpose, class, activity, or real-world meaning between the two sets of images. \n\nRule on framing the target property:\nThe target property must be the object type itself (e.g., rivers vs. roads) or a direct equivalent (e.g., “presence of water vs. absence of water”).\n\nAn answer is incorrect if it:\n\n· Replaces the object types with a broader category (e.g., “natural landscapes vs. human-made infrastructure”, “rural vs. urban”, “ecologically rich vs. degraded”).\n· Adds an unnecessary distinguishing attribute that is not guaranteed by the reference answer (e.g., “winding rivers” — winding is typical but not required; “urban roads” — roads can be rural).\n· Frames the distinction primarily through secondary features (e.g., vegetation, color, dryness, shadow of planes) instead of naming the objects.\n\nPurely descriptive adjectives (e.g., “wide brown river”, “narrow gray road”) are allowed as long as the core distinction remains “river vs. road” and the adjectives are not presented as universal class properties.\nIf reference answer includes two distinct classes, it is enough for user to correctly pinpoint one of them for the answer to be considered correct.'"
+  }
+}

LLM-as-a-Judge /tasks_en.json

@@ -0,0 +1,614 @@
+{
+  "tasks": [
+    {
+      "name": "bb_m_01.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_02.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_03.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_04.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_08.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_09.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_10.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_11.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_12.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_13.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_14.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_15.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_18.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_19.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_20.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_22.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_23.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_24.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_27.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_28.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_29.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_30.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_32.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_33.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_36.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_37.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_38.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_41.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_45.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_46.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_r_48.png",
+      "type": "size",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_49.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_51.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_53.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_54.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_56.png",
+      "type": "shape",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_57.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_58.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_59.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_60.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_61.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_62.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_63.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_65.png",
+      "type": "same",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_66.png",
+      "type": "same",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_67.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_68.png",
+      "type": "number",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_69.png",
+      "type": "number",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_70.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_71.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_73.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_74.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_75.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_76.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_79.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_80.png",
+      "type": "semantic",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_83.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_87.png",
+      "type": "semantic",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_89.png",
+      "type": "shape",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_91.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_92.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_94.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_95.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_96.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_97.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_98.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_r_99.png",
+      "type": "presence",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_100.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_101.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_102.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_103.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_104.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_105.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_106.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_107.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_108.png",
+      "type": "shape",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_109.png",
+      "type": "semantic",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_110.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_111.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_112.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_113.png",
+      "type": "shape",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_114.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_115.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_116.png",
+      "type": "size",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_117.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_118.png",
+      "type": "number",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_120.png",
+      "type": "same",
+      "complexity": 3
+    },
+    {
+      "name": "bb_m_121.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_122.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_m_124.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_125.png",
+      "type": "same",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_05.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_08.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_11.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_12.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_14.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_15.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_17.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_18.png",
+      "type": "spatial",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_19.png",
+      "type": "presence",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_20.png",
+      "type": "spatial",
+      "complexity": 3
+    },
+    {
+      "name": "bb_s_22.png",
+      "type": "size",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_23.png",
+      "type": "number",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_25.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_28.png",
+      "type": "size",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_29.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_30.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_34.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_35.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_r_38.png",
+      "type": "semantic",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_39.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_41.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_42.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_43.png",
+      "type": "semantic",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_44.png",
+      "type": "number",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_49.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_53.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_55.png",
+      "type": "presence",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_57.png",
+      "type": "spatial",
+      "complexity": 2
+    },
+    {
+      "name": "bb_s_58.png",
+      "type": "number",
+      "complexity": 1
+    },
+    {
+      "name": "bb_s_59.png",
+      "type": "size",
+      "complexity": 1
+    },
+    {
+      "name": "bb_m_126.png",
+      "type": "semantic",
+      "complexity": 1
+    }
+  ]
+}

prompting_strategies.json

@@ -0,0 +1,43 @@
+{
+    "strategies": [
+        {
+            "strategy": "direct",
+            "prompts": [
+                "Here is the Bongard problem collage. Provide your answer as two plain sentences: first the concept for all left images, then the concept for all right images."
+            ]
+        },
+        {
+            "strategy": "descriptive-direct",
+            "prompts": [
+                "This is a single image from either the left or right side of a Bongard problem. Do not solve the problem yet. Describe this image in detail. Focus on all visual features: objects, shapes, colors, spatial arrangements, textures, and any other distinctive properties. Be thorough, as these descriptions will later be used to identify a common concept.",
+                "Here is the Bongard problem collage. Below are the detailed descriptions of each left image and each right image. Based on this, provide your answer. Left class images descriptions: {}. Right class images descriptions: {}."
+            ]
+        },
+        {
+            "strategy": "contrastive-direct",
+            "prompts": [
+                "Here is a pair: one image from the left side of a Bongard problem and one from the right side. Do not solve the full problem yet. Instead, carefully examine both images and list all differences between them. Focus on contrasting visual features: objects, shapes, colors, positions, textures, counts, orientations, or any other properties that distinguish the left image from the right image. Output your answer as a simple list of differences. Use plain text, no markdown.",
+                "Here is the full Bongard problem collage. Below are all the difference lists obtained from comparing left-right image pairs. Based on this information, determine the common concept that unites all left images but no right images, and the common concept that unites all right images but no left images. Provide your answer."
+            ]
+        },
+        {
+            "strategy": "contrastive-iterative",
+            "prompts": [
+                "Here is a pair: one image from the left side of a Bongard problem and one from the right side. Сarefully examine both images and list all differences between them. Focus on contrasting visual features: objects, shapes, colors, positions, textures, counts, orientations, or any other properties that distinguish the left image from the right image. Propose a candidate concept that distinguishes the left image from the right image.",
+                "Here is the next left-right pair. Your goal is to generalize concept to fit all of the pairs you have seen. Carefully examine new pair and find all differences between images. Focus on contrasting visual features: objects, shapes, colors, positions, textures, counts, orientations, or any other properties that distinguish the left image from the right image. Your previous candidate concept is: {}. Check if the previous concept applies for new pair. If fully correct, output the same concept unchanged. If partially correct (some parts hold, some do not), refine the concept by removing or adjusting the failing parts, keeping only the aspects valid for all left images seen so far and false for all right images. If completely wrong, discard it and formulate a completely new concept based on all pairs seen now (including this one). Output the updated concept as one detailed sentence covering all distinguishing features common to all processed pairs.",
+                "This is the last left-right pair. Based on all six pairs and your iterative refinement, provide your final answer. Output exactly two plain sentences: the first describes the concept that holds for all left images but no right images, the second describes the concept that holds for all right images but no left images. Your iterative concept is: {}."
+            ]
+        },
+        {
+            "strategy": "descriptive-iterative",
+            "prompts": [
+                "This is the first image from one side of a Bongard problem. Later you will see more images from this same side. Your goal is to formulate a general description of all images on this side for solving the problem later. Do not solve the problem yet. Describe this image in detail. Focus on all visual features: objects, shapes, colors, spatial arrangements, textures, counts, orientations, and any other distinctive properties. Be thorough, as these descriptions will later be used to identify a common concept.",
+                "Here is the next image from the same side. Your goal is to generalize the class description to fit all images from this side that you have seen so far. Carefully examine this new image and list all its relevant visual features: objects, shapes, colors, spatial arrangements, textures, counts, orientations, or any other distinctive properties. Your previous candidate description for this side is: {}. Check if is applies to the new image. If fully correct, output the same description unchanged. If partially correct (some parts hold, some do not), refine the description by removing or adjusting the failing parts, keeping only what is true for all images seen so far. If completely wrong, discard it and formulate a completely new description based on all images seen now (including this one).",
+                "This is the last image from this side. Based on all six images and your iterative refinement, provide the final description that unites all images on this side. Your iterative description is: [previous concept].",
+                "Here is the Bongard problem collage. Below are generalized descriptions of left and right sides of images. Based on this, solve Bongard problem. Left class images descriptions: {}. Right class images descriptions: {}."
+            ]
+        }
+    ],
+    "dataset": "./BMRS/",
+    "model": "MODEL_NAME"
+}