{
 "built_at": "2026-09-24T21:29:39Z",
 "snapshot": "2025-07-28",
 "graph_read_at": "2026-09-24T18:15:14+00:00",
 "limit": null,
 "catalog_counts": {
  "leaderboards": 11570,
  "leaderboards_with_rows": 11026,
  "tasks": 5042,
  "datasets": 12178,
  "methods": 2290,
  "papers": 625187,
  "papers_with_code": 218852,
  "code_links": 300161,
  "lb_rows": 76200,
  "graph_papers": 54410,
  "graph_ran": 38943
 },
 "arxiv_metadata": {
  "present": true,
  "licence": "arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)",
  "file_sha256_as_read": "ce2880e03d9a9b7d2f3864b90677793c79ba1d63a3db4ed82251ced94a2aaecc",
  "papers": 9801,
  "counts": {
   "from_arxiv": 7670,
   "from_arxiv_daily": 287,
   "from_arxiv_api": 1254,
   "from_arxiv_backlog": 590
  },
  "created_at": "2026-09-24T16:22:16+00:00"
 },
 "extracted_results": {
  "present": false,
  "source": "lb_extract (leaderboard-shaped extractor), verified",
  "file": "cwp/data/lb_rows_run_verified.jsonl",
  "file_sha256_as_read": null,
  "file_bytes_as_read": 0,
  "file_lines": {},
  "papers_in_file": 0,
  "status": {},
  "boards_proposed": 0,
  "refused_by_rule": {},
  "refused_by_rule_cells": 0,
  "verifier": {},
  "declined_by_site": {},
  "entries_accepted": 0,
  "cells_accepted": 0,
  "boards_with_entries": 0,
  "papers_with_entries": 0,
  "configuration_not_shown": {},
  "measurement": {
   "test_papers": 883,
   "papers_with_output": 881,
   "papers_with_output_from": null,
   "accepted_entries": 110,
   "judged_true": 108,
   "judged": 110,
   "wilson95_lower": 0.9361,
   "passed": true,
   "frozen_commit": "0e3de0df94",
   "result_commit": null,
   "measured_on": "2026-09-24"
  },
  "measurement_sha256_as_read": "f33f040865aa9b360a85c26e097b3f2d66cb34149574b9366039d37791554b75"
 },
 "pages": {
  "home": {
   "per_page": 30,
   "per_stream_per_page": 15,
   "cards_cap_per_stream": 150,
   "since_archive_tile": null,
   "sample_gap": {
    "samples": 294260,
    "samples_ran": 127244,
    "samples_unverified": 167016,
    "papers_harvested": 51887,
    "papers_ran": 38943,
    "papers_harvested_nothing_ran": 12944
   },
   "ran": {
    "pages": 10,
    "post_freeze": {
     "cards": 150,
     "total": 5105,
     "unplaceable": 0,
     "with_arxiv_metadata": 5105
    },
    "archive": {
     "cards": 150,
     "total": 31700,
     "unplaceable": 1,
     "dated_after_snapshot": 0,
     "placed_by_id_month": 819,
     "max_archive_date": "2025-07-17",
     "code_no_paper_row": 1
    }
   },
   "latest": {
    "pages": 10,
    "post_freeze": {
     "cards": 150,
     "total": 9581,
     "unplaceable": 0,
     "with_arxiv_metadata": 9280
    },
    "archive": {
     "cards": 150,
     "total": 218469,
     "unplaceable": 382,
     "dated_after_snapshot": 4,
     "placed_by_id_month": 49179,
     "max_archive_date": "2025-09-24",
     "code_no_paper_row": 1
    }
   }
  },
  "paper": {
   "paper_pages": 225441,
   "post_freeze_pages": 9581,
   "links_only_arxiv_metadata": {
    "source": "arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license), from the Kaggle arXiv metadata snapshot of 2026-09-12",
    "rows_links_only_with_page": 49561,
    "with_arxiv_id": 49540,
    "filled": 48630,
    "not_in_snapshot": 17,
    "title_disagrees": 893
   },
   "post_freeze_pages_with_arxiv_metadata": 9280,
   "post_freeze_pages_without_arxiv_metadata": 301,
   "post_freeze_pages_without_repository_for_attached_rows": 0,
   "pages_with_extracted_results_section": 0,
   "post_freeze_pages_in_catalog": 9581,
   "papers_in_catalog": 625187,
   "spam_title_withheld": {
    "archive_pages": {
     "count": 0,
     "ids": []
    },
    "post_freeze_pages": {
     "count": 0,
     "ids": []
    }
   },
   "not_shown_abstract_only": 399746,
   "syntology_links": {
    "paper_pages_with_atlas_link": 110823,
    "paper_pages_with_agent_route": 49446,
    "paper_pages": 235022,
    "by_page_type": {
     "paper_pages": 225441,
     "paper_pages_arxiv_id_not_in_atlas": 109884,
     "paper_pages_with_agent_route": 42388,
     "paper_pages_with_atlas_link": 102659,
     "paper_post_freeze_pages": 9581,
     "paper_post_freeze_pages_arxiv_id_not_in_atlas": 1417,
     "paper_post_freeze_pages_with_agent_route": 7058,
     "paper_post_freeze_pages_with_atlas_link": 8164
    },
    "developers": "https://syntology.ai/developers",
    "atlas_base": "https://app.syntology.ai/",
    "pages_atlas_link_withheld_unrecognised_id": 0,
    "rule": "Atlas link only for an arXiv id (new-style NNNN.NNNNN or old-style archive/NNNNNNN) in the Atlas id list read by this build; an id in the list of any other shape is withheld and counted; get_harvested_code_for_paper only for a paper with an arXiv id and harvested samples; get_code on every harvested sample row, by its code_sha256"
   }
  },
  "sota": {
   "sota_pages": 1,
   "area_pages": 16,
   "categories": 16,
   "benchmarks_with_rows": 11026,
   "not_shown_lb_tables_without_rows": 544,
   "not_shown_lb_tables_without_task_page": 24,
   "not_shown_tasks_without_category": 1212,
   "tasks_without_category_but_with_benchmarks": 20,
   "graph_papers_without_archive_code_link": 926,
   "graph_papers_without_archive_row": 12022,
   "graph_papers_without_archive_row_post_freeze": 9581
  },
  "leaderboard": {
   "leaderboard_pages": 11570,
   "empty_tables": 544,
   "rows_rendered": 76200,
   "metric_columns_direction_inferred": 14397,
   "metric_columns_direction_unknown": 8506,
   "tables_without_task": 24,
   "tables_placeholder_categories": 135,
   "tables_with_since_archive_rows": 0,
   "since_archive_rows_rendered": 0,
   "leaderboards_in_catalog": 11570,
   "lb_rows_in_catalog": 76200,
   "content_hygiene": {
    "rules": "cwp/cwp_lib/data.py PROSE_RULES / NAME_RULES (NFKC-folded text); applied at build time to TEMP copies of the catalog tables, never to the catalog file",
    "withheld": {
     "datasets": {
      "n": 36,
      "by_reason": {
       "travel_booking": 7,
       "contact_centre": 1,
       "affiliate": 24,
       "gambling": 4
      }
     },
     "methods": {
      "n": 31,
      "by_reason": {
       "travel_booking": 7,
       "gambling": 3,
       "pharma": 2,
       "contact_centre": 10,
       "test_entry": 2,
       "affiliate": 6,
       "name_and_text_replaced_by_affiliate": 1
      }
     },
     "tasks": {
      "n": 2,
      "by_reason": {
       "contact_centre": 2
      }
     },
     "leaderboards": {
      "n": 27,
      "by_reason": {
       "task_withheld": 1,
       "only_row_was_a_test_entry": 23,
       "every_row_withheld_as_spam": 3
      }
     },
     "lb_rows": {
      "n": 26,
      "by_reason": {
       "one_row_on_dummy_anchor_paper": 23,
       "contact_centre": 2,
       "travel_booking": 1
      }
     }
    },
    "withheld_slugs": {
     "datasets": [
      "100-guide-expedia-booking-support",
      "100-guidelines-expedia-contact-number-for",
      "100-guidlines-what-number-do-i-call-to-book-a-3",
      "55-club-invite-code-385337634526-get-instant",
      "55-club-invite-code-385337634526-use-this-get",
      "91-club-invite-code-4325118822517-use-this",
      "can-i-earn-frequent-flyer-miles-with-expedia",
      "clearcachehelp",
      "gemini-wallet-support-9-official-methods",
      "how-to-login-multiple-account",
      "join-tiranga-today-using-tiranga-game-invite",
      "join-tiranga-with-invite-code-5882818674566",
      "lottery-7-invite-code",
      "lottery-7-invite-code-8254516456889",
      "lottery-7-invite-code-8254516456889-use-this",
      "ok-win-invite-code",
      "ok-win-invite-code-8522713375003-get-instant",
      "ok-win-invite-code-8522713375003-register-now",
      "ok-win-invite-code-8522713375003-register-now-1",
      "ok-win-invite-code-8522713375003-unlock",
      "ok-win-invite-code-8522713375003-use-this",
      "raja-game-register-login-2025",
      "register-on-tiranga-win-big-with-exclusive",
      "step-by-step-apply-55-club-invite-code",
      "step-by-step-apply-91-club-invite-code",
      "step-by-step-apply-bdg-win-invite-code",
      "step-by-step-apply-daman-invite-code",
      "step-by-step-apply-ok-win-invite-code",
      "step-by-step-apply-tiranga-invite-code",
      "tiranga-game-invite-code-5882818674566-join",
      "tiranga-game-invite-code-get-rewards-start",
      "tiranga-invite-code-5882818674566-use-this",
      "tiranga-invite-code-5882818674566-your-key-to",
      "trust-wallet-support-how-to-use-1-888-416",
      "use-5882818674566-on-tiranga-app-to-claim",
      "what-is-91-club-invite-code-4325118822517"
     ],
     "methods": [
      "19-ways-to-get-assistance-from-etihad-full",
      "25th-wedding-anniversary-gifts",
      "6-ways-immediate-solutions-to-connect-with",
      "air-canada-fee-150",
      "alpha-bio-cbd-gummies-for-stress-anxiety-what",
      "cheapest-air-canada",
      "d365bc",
      "easy-way-to-login-cox-webmail-account",
      "effortless-microsoft-support-get-help-when",
      "golden-queue-managers",
      "gws-to-o365-migration",
      "is-the-cheapest-to-buy-southwest",
      "kamagra-100",
      "life2home",
      "microsoft-support-fast-reliable-help-for-all",
      "microsoft-support-your-guide-to-seamless",
      "navigating-microsoft-help-fast-friendly-and",
      "o365-to-o365",
      "poker-software-developer",
      "primafelicitas",
      "reddy-book-club",
      "sadasd",
      "simple-methods-for-canon-printer-is-offline",
      "test-method",
      "the-problem-six-ways-to-contact-to-someone-at",
      "the-problem-six-ways-to-contact-to-someone-at-2",
      "the-week-to-fly-southwest",
      "to-book-on-southwest",
      "uickbooks-payroll-support",
      "w88club",
      "week-to-fly-southwest"
     ],
     "tasks": [
      "how-to-refund-a-wrong-transaction-in-phonepe",
      "how-to-refund-a-wrong-transaction-in-phonepe-2"
     ],
     "leaderboards": [
      "10-shot-image-generation-on-1",
      "10-shot-image-generation-on-flyingthings3d",
      "10-shot-image-generation-on-music21-1",
      "3d-anomaly-detection-and-segmentation-on-1",
      "3d-face-alignment-on-facewarehouse",
      "3d-facial-expression-recognition-on",
      "articles-on",
      "articles-on-1",
      "deepfake-detection-on-1",
      "deepfake-detection-on-2",
      "deepfake-detection-on-wiiekh-chmkhlip-ai-ch",
      "face-swapping-on-1",
      "how-to-refund-a-wrong-transaction-in-phonepe",
      "image-restoration-on",
      "license-plate-detection-on-common-voice",
      "object-detection-on-01-01-19679682867",
      "object-detection-on-100-sleep-nights-of-8",
      "object-detection-on-10000-people-human-pose",
      "on-scannet",
      "on-vgg-face",
      "semantic-segmentation-on-uyfds",
      "shadow-removal-on",
      "style-transfer-on-01-01-1967-and-2-3-8-6-8-1",
      "style-transfer-on-9",
      "unconditional-video-generation-on",
      "video-object-detection-on-01-01-19679682867",
      "weakly-supervised-semantic-segmentation-on-24"
     ]
    },
    "board_rows_withheld_board_kept": {},
    "fields_blanked": {
     "datasets.license_url": 1,
     "datasets.description": 1,
     "datasets.license_name": 1,
     "methods.description": 3,
     "tasks.description": 1
    },
    "dataset_names_restored": {
     "adience": {
      "from": "mebeblurf",
      "to": "Adience"
     }
    },
    "paper_tag_names_withheld_distinct": {
     "methods": 255,
     "tasks": 1
    },
    "method_links_shown_as_plain_text": {
     "see_code_not_a_code_host": 26,
     "source_not_a_paper_host": 12
    },
    "counts": {
     "datasets_license_url_blanked:spam_host": 1,
     "datasets_description_blanked:darknet": 1,
     "datasets_name_restored_from_slug": 1,
     "datasets_license_name_blanked:vandal_name_on_restored_row": 1,
     "datasets_withheld:travel_booking": 7,
     "datasets_withheld:contact_centre": 1,
     "datasets_withheld:affiliate": 24,
     "datasets_withheld:gambling": 4,
     "methods_description_blanked:phone_number": 1,
     "methods_withheld:travel_booking": 7,
     "methods_withheld:gambling": 3,
     "methods_withheld:pharma": 2,
     "methods_withheld:contact_centre": 10,
     "methods_description_blanked:junk_token": 2,
     "methods_withheld:test_entry": 2,
     "methods_withheld:affiliate": 6,
     "methods_withheld:name_and_text_replaced_by_affiliate": 1,
     "tasks_description_blanked:phone_number": 1,
     "tasks_withheld:contact_centre": 2,
     "leaderboards_withheld:task_withheld": 1,
     "lb_rows_withheld:one_row_on_dummy_anchor_paper": 23,
     "lb_rows_withheld:contact_centre": 2,
     "lb_rows_withheld:travel_booking": 1,
     "leaderboards_withheld:only_row_was_a_test_entry": 23,
     "leaderboards_withheld:every_row_withheld_as_spam": 3,
     "paper_methods_tag_names_withheld_distinct": 255,
     "paper_tasks_tag_names_withheld_distinct": 1,
     "method_source_link_not_a_paper_host": 12,
     "method_code_link_not_a_code_host": 26
    }
   }
  },
  "task": {
   "task_pages": 5042,
   "tasks_in_catalog": 5042,
   "with_description": 2114,
   "empty_tasks_rendered": 21,
   "tasks_with_leaderboard_tables": 2090,
   "tasks_with_benchmarks": 2084,
   "tasks_with_datasets": 3715,
   "dataset_task_refs_without_task_page": 2,
   "papers_shown_per_task": 30,
   "paper_lists_code_only": true,
   "paper_lists_shorter_than_cap": 3870,
   "description_links_kept": 99,
   "description_links_unwrapped_to_text": 0,
   "description_bare_pwc_urls_linked": 2,
   "description_relative_images_dropped": 0
  },
  "dataset": {
   "dataset_pages": 12172,
   "withheld_stub_pages": 6,
   "index_pages": 254,
   "modality_pages": 253,
   "task_pages": 3807,
   "language_pages": 486,
   "modalities": 39,
   "task_facets": 3717,
   "language_facets": 367,
   "datasets_in_archive": 12214,
   "datasets_listed": 12172,
   "rows_withheld_by_content_rules": 36,
   "descriptions_blanked_by_content_rules": 1,
   "names_restored_by_content_rules": 1,
   "rows_withheld": 6,
   "titles_withheld_as_spam": 6,
   "bare_rows_withheld": 0,
   "anchor_exempt": 0,
   "descriptions_withheld_as_spam": 1,
   "withheld_max_num_papers": 1,
   "withheld_with_benchmarks": 0,
   "description_links_unlinked_dead_slug": 5,
   "datasets_with_benchmarks": 3780,
   "datasets_with_a_paper_row": 3741,
   "intro_paper_unresolved": 158
  },
  "method": {
   "methods_index": 1,
   "area_pages": 7,
   "category_pages": 352,
   "category_alias_pages": 8,
   "method_pages": 2290,
   "methods_in_catalog": 2290,
   "methods_rendered_population": 2290,
   "methods_excluded": {},
   "collection_pairs_in_catalog": 352,
   "collection_pairs_without_area": 2,
   "methods_without_collection": 12,
   "pwc_category_aliases_served": [
    "attention-mechanisms-1",
    "discrimination",
    "factorized-attention",
    "pooling-operation",
    "q-learning",
    "segmentation-models",
    "speech-separation",
    "vision-transformer"
   ],
   "pwc_category_aliases_unresolved": [
    "face-to-face-translation",
    "image-dataset-comparison-metric",
    "image-manipulation-models",
    "interactive-semantic-segmentation-models",
    "medical-image-deblurring",
    "monocular-depth-estimation-models",
    "oriented-object-detection-models",
    "super-resolution-models",
    "video-inpainting-models"
   ],
   "description_states": {
    "present": 1987,
    "absent": 262,
    "withheld_vandalized": 6,
    "placeholder": 35
   },
   "full_name_withheld_page_gate": 0,
   "source_or_code_urls_not_http_dropped": 10,
   "archive_links_not_linked": {
    "source_shown_as_text": 12,
    "source_dropped_spam_host": 0,
    "code_shown_as_text": 14,
    "code_dropped_spam_host": 2
   },
   "content_hygiene": {
    "n": 31,
    "by_reason": {
     "travel_booking": 7,
     "gambling": 3,
     "pharma": 2,
     "contact_centre": 10,
     "test_entry": 2,
     "affiliate": 6,
     "name_and_text_replaced_by_affiliate": 1
    }
   }
  },
  "code": {
   "index_pages": 1,
   "entry_pages": 7868,
   "entries_eligible": 7868,
   "min_papers": 5,
   "distinct_entries": 179672,
   "papers_with_samples": 51887,
   "samples": 294260,
   "samples_without_entry": 0,
   "slug_collision_groups": 213,
   "slugs_suffixed": 228,
   "case_only_name_pairs": 194,
   "names_with_exactly_k_papers": {
    "1": 146890,
    "2": 16565,
    "3": 5551,
    "4": 2798,
    "5": 1716,
    "6": 1046,
    "8": 558,
    "10": 384,
    "20": 80,
    "50": 10,
    "100": 3
   },
   "names_with_at_least_k_papers": {
    "2": 32782,
    "3": 16217,
    "4": 10666,
    "5": 7868,
    "6": 6152,
    "8": 4270,
    "10": 3262
   },
   "search_code_rows": 7868,
   "search_code_bytes": 370723,
   "seconds": 10.9
  },
  "census": {
   "index_pages": 1,
   "name_pages": 296,
   "names_eligible": 296,
   "names_split": 192,
   "min_papers": 5,
   "papers_covered": 2630,
   "implementations_compared": 2199,
   "clusters": 1187,
   "buckets": 670,
   "bucket_fields": [
    "rank",
    "kind",
    "dtype"
   ],
   "census_dir": "/Users/qed/Documents/syntology-final/groundwork/55_differential_census/census/v2_2026-09-22",
   "files": 17,
   "lines": 9976,
   "duplicate_lines": 0,
   "samples": 9916,
   "compared": 4946,
   "torch_reference_rows_not_counted": 60,
   "not_run_shared": 2689,
   "own_args_only": 2272,
   "own_args_same_as_fixture": 2225,
   "own_args_on_pages": 153,
   "not_digested": 9,
   "not_digested_types": {
    "ndarray": 2,
    "NoneType": 2,
    "Tensor": 5
   },
   "not_digested_on_pages": 0,
   "names_in_files": 3320,
   "papers_in_files": 7537,
   "class_bearing_samples": 1808,
   "class_bearing_ran_shared": 586,
   "withdrawn_key_in_export": false,
   "withdrawn_excluded": 0,
   "census_samples_found_in_graph_export": 9916,
   "absence": {
    "held_not_in_census": 5910,
    "never_ran_in_files": 781,
    "never_compared": 675,
    "thin": 1568,
    "eligible_only_via_own": 133,
    "own_args_only": 2272,
    "own_args_same_as_fixture": 2225,
    "own_on_pages": 153,
    "not_digested": 9,
    "not_digested_types": [
     [
      "NoneType",
      2
     ],
     [
      "Tensor",
      5
     ],
     [
      "ndarray",
      2
     ]
    ],
    "undig_on_pages": 0,
    "census_names_not_held": 0,
    "class_bearing_samples": 1808,
    "class_bearing_ran_shared": 586,
    "class_bearing_names_with_page": 24
   },
   "census_dates": [
    "2026-09-22"
   ],
   "battery_sha256": [
    "b986f7e04d794a0d88ad4c5f32cf63ec3590b5deff0192150737bc5f1c0b4677"
   ],
   "runner_sha256": [
    "5a452d0e7c0da5b80771d1be2afe3572e253568cf5c5d59a0d08ebd663d00808"
   ],
   "sidecars_read": 17,
   "sidecars_missing": 0,
   "slug_collision_groups": 9,
   "slugs_suffixed": 9,
   "seconds": 5.5
  },
  "search": {
   "index_rows": 265558,
   "paper_rows": 235022,
   "words": 91964,
   "buckets": 11228,
   "bucket_files": 1001,
   "stop_only_rows": 5,
   "largest_files_rows": [
    2500,
    2500,
    2500
   ],
   "row_copies": 2021888,
   "index_bytes": 212972081,
   "manifest_bytes": 239875,
   "paper_slugs_derived_client_side": 199003
  },
  "static_pages": {
   "pages": 6
  },
  "contribute": {
   "pages": 2,
   "participation": true,
   "corrections_public": {
    "present": false,
    "by_decision": {},
    "items": 0
   },
   "overrides_applied": 0,
   "overrides_not_applied": 0
  }
 },
 "store": {
  "pages": 553786,
  "shards": 4096,
  "by_type": {
   "home": 20,
   "home_json": 20,
   "paper": 225441,
   "paper_json": 235022,
   "paper_post_freeze": 9581,
   "sota": 1,
   "sota_json": 1,
   "area": 16,
   "area_json": 16,
   "leaderboard": 11570,
   "leaderboard_json": 11570,
   "task": 5042,
   "task_json": 5042,
   "datasets_index": 254,
   "datasets_modality": 253,
   "datasets_task": 3807,
   "datasets_language": 486,
   "datasets_withheld_json": 1,
   "dataset": 12172,
   "dataset_json": 12178,
   "dataset_withheld": 6,
   "methods_index": 1,
   "method_area": 7,
   "method_category": 352,
   "method_category_alias": 8,
   "method": 2290,
   "method_json": 2290,
   "code_index": 1,
   "code_entry": 7868,
   "code_json": 7868,
   "census": 297,
   "census_json": 296,
   "search": 1,
   "static": 6,
   "contribute": 1,
   "corrections": 1
  },
  "bytes_by_type": {
   "home": 918911,
   "home_json": 434398,
   "paper": 2877976053,
   "paper_json": 978450916,
   "paper_post_freeze": 183919023,
   "sota": 29019,
   "sota_json": 13023,
   "area": 1365790,
   "area_json": 735391,
   "leaderboard": 216875098,
   "leaderboard_json": 62895477,
   "task": 91961859,
   "task_json": 36726168,
   "datasets_index": 32226322,
   "datasets_modality": 11985624,
   "datasets_task": 70557314,
   "datasets_language": 12128614,
   "datasets_withheld_json": 4092,
   "dataset": 145289589,
   "dataset_json": 42174200,
   "dataset_withheld": 42972,
   "methods_index": 15266,
   "method_area": 292498,
   "method_category": 2817977,
   "method_category_alias": 50814,
   "method": 32879944,
   "method_json": 11888890,
   "code_index": 23151,
   "code_entry": 181075111,
   "code_json": 109521085,
   "census": 4759670,
   "census_json": 3268125,
   "search": 6035,
   "static": 80857,
   "contribute": 12399,
   "corrections": 5822
  },
  "bytes_total": 5113407497
 },
 "sitemap_urls": 279467,
 "arxiv_map_entries": 222124,
 "redirects": {
  "redirects": 10374,
  "trailing_slash": 290,
  "legacy_task_subpages": 10084
 },
 "download": {
  "graph_layer_status_rewritten": 53454,
  "graph_layer.jsonl": 15835250,
  "repo_status.json": 1313180
 },
 "unlinked_internal_links_by_type": {
  "dataset": 7,
  "paper": 25,
  "task": 6,
  "method": 218
 },
 "syntology_links": {
  "atlas": {
   "url": "https://app.syntology.ai/data/ids.json",
   "fetched_at": "2026-09-24T21:29:46Z",
   "status": 200,
   "bytes": 9448420,
   "sha256": "92f118c4f9f0ca6296b84f76fbf33b03868ce3c0f1ba239e159e08caad421bb3",
   "keys": 528367,
   "error": null,
   "enabled": true,
   "note": "Atlas links only for arXiv ids in this list",
   "mode": "auto"
  },
  "paper_pages_with_atlas_link": 110823,
  "paper_pages_with_agent_route": 49446,
  "paper_pages": 235022,
  "by_page_type": {
   "paper_pages": 225441,
   "paper_pages_arxiv_id_not_in_atlas": 109884,
   "paper_pages_with_agent_route": 42388,
   "paper_pages_with_atlas_link": 102659,
   "paper_post_freeze_pages": 9581,
   "paper_post_freeze_pages_arxiv_id_not_in_atlas": 1417,
   "paper_post_freeze_pages_with_agent_route": 7058,
   "paper_post_freeze_pages_with_atlas_link": 8164
  },
  "developers": "https://syntology.ai/developers",
  "atlas_base": "https://app.syntology.ai/",
  "pages_atlas_link_withheld_unrecognised_id": 0,
  "rule": "Atlas link only for an arXiv id (new-style NNNN.NNNNN or old-style archive/NNNNNNN) in the Atlas id list read by this build; an id in the list of any other shape is withheld and counted; get_harvested_code_for_paper only for a paper with an arXiv id and harvested samples; get_code on every harvested sample row, by its code_sha256"
 },
 "participation": {
  "mode": "auto",
  "enabled": true,
  "probe": {
   "endpoint": "https://syntology.ai/api/cwp/report",
   "origin": "https://codewithpapers.app",
   "at": "2026-09-24T21:29:45Z",
   "status": 204,
   "allow_origin": "https://codewithpapers.app",
   "error": null,
   "live": true
  },
  "csp_probe": {
   "url": "https://codewithpapers.app/",
   "need_in_connect_src": "https://syntology.ai",
   "at": "2026-09-24T21:29:45Z",
   "status": 200,
   "connect_src": "'self' https://syntology.ai",
   "error": null,
   "allows": true
  },
  "note": "endpoint live and live CSP allows it",
  "has_public_decisions": false,
  "report_links_by_item_type": {
   "code_link": 311823,
   "page": 274267,
   "sample": 343515,
   "archive_row": 209999
  },
  "report_links_skipped_unservable_path": {}
 },
 "overrides": {
  "present": false,
  "entries": 0,
  "applied": 0,
  "not_applied": 0,
  "by_kind": {
   "code_links_add": {},
   "code_links_remove": {},
   "archive_values": {},
   "titles": {},
   "suppress_syntology_rows": {},
   "suppress_samples": {},
   "notes": {}
  },
  "results": [],
  "removed": [],
  "notes_rendered_on_pages": 0,
  "views": {
   "installed": false,
   "titles": 0,
   "values_rows": 0,
   "links_removed_papers": 0,
   "links_added_papers": 0
  }
 },
 "overrides_verify": {
  "skipped": "no override applied"
 },
 "not_shown": {
  "abstract_only_papers_without_a_page": 399746,
  "papers_withheld_spam_title": {
   "archive_pages": {
    "count": 0,
    "ids": []
   },
   "post_freeze_pages": {
    "count": 0,
    "ids": []
   }
  },
  "catalog_report": {
   "code_links": 300161,
   "collection_description_vandalized_blanked": 65,
   "datasets": 12214,
   "datasets_vandalized_text_blanked": 201,
   "lb_dataset_node_without_table": 190,
   "lb_dataset_unresolved": 3096,
   "lb_rows": 76227,
   "lb_rows_paper_slug_without_page": 0,
   "lb_rows_paper_url_not_pwc": 72508,
   "lb_rows_resolved_by_none": 3740,
   "lb_rows_resolved_by_url_abs": 72487,
   "lb_rows_without_paper_url": 3719,
   "lb_task_unresolved": 26,
   "leaderboards": 11597,
   "method_slug_collision_suffixed": 2,
   "methods": 2321,
   "methods_embedded_distinct_names": 14827,
   "methods_from_embedded": 441,
   "methods_from_table": 1880,
   "methods_vandalized_text_blanked": 22,
   "paper_method_name_spam_rows": 37766,
   "paper_methods_unresolved_names": 386,
   "paper_methods_unresolved_rows": 4230,
   "paper_tasks_unresolved_names": 246,
   "paper_tasks_unresolved_rows": 29815,
   "papers": 625187,
   "papers_from_abstracts": 575626,
   "papers_links_only": 49561,
   "spam_anchor_titles": 5,
   "task_node_unnamed": 1,
   "task_nodes": 6448,
   "task_slug_collision_suffixed": 10,
   "task_slug_derived": 1324,
   "task_slugs_from_datasets": 3720,
   "tasks": 5044,
   "tasks_only_in_datasets_table": 1069,
   "tasks_vandalized_text_blanked": 5,
   "unresolved_task_names_top": [
    [
     "GPU",
     5629
    ],
    [
     "Position",
     3684
    ],
    [
     "valid",
     3589
    ],
    [
     "Relation",
     3487
    ],
    [
     "Generative Adversarial Network",
     3110
    ],
    [
     "counterfactual",
     2765
    ],
    [
     "CPU",
     2231
    ],
    [
     "Triplet",
     1626
    ],
    [
     "Functional Connectivity",
     460
    ],
    [
     "zero-shot-classification",
     422
    ],
    [
     "Hippocampus",
     300
    ],
    [
     "8k",
     202
    ],
    [
     "Image Description",
     154
    ],
    [
     "Medical Question Answering",
     139
    ],
    [
     "mbpp",
     129
    ],
    [
     "Avg",
     117
    ],
    [
     "dialog state tracking",
     113
    ],
    [
     "MedQA",
     80
    ],
    [
     "CoLA",
     78
    ],
    [
     "SST-2",
     66
    ],
    [
     "AudioCaps",
     64
    ],
    [
     "Grapheme-to-Phoneme Conversion",
     62
    ],
    [
     "Causal Language Modeling",
     52
    ],
    [
     "TextVQA",
     47
    ],
    [
     "Pathfinder",
     43
    ]
   ],
   "unresolved_method_names_top": [
    [
     "Library",
     1475
    ],
    [
     "Travel",
     1089
    ],
    [
     "American",
     835
    ],
    [
     "None",
     244
    ],
    [
     "Uphold",
     114
    ],
    [
     "Michael",
     15
    ],
    [
     "(USA Guide)",
     12
    ],
    [
     "Contact Information",
     12
    ],
    [
     "RSM",
     11
    ],
    [
     "Identify the Issue",
     10
    ],
    [
     "hello",
     10
    ],
    [
     "Real Person",
     7
    ],
    [
     "Quickbooks payroll support",
     5
    ],
    [
     "Roadrunner",
     5
    ],
    [
     "Quickbooks",
     4
    ],
    [
     "DECA",
     4
    ],
    [
     "tst",
     4
    ],
    [
     "TRON",
     3
    ],
    [
     "Trust Wallet",
     2
    ],
    [
     "Report a Problem",
     2
    ],
    [
     "Human in Seconds",
     2
    ],
    [
     "QuickBooks Enterprise Support Experts and Fix Issues Fast",
     1
    ],
    [
     "You Can Reach Metamask Anytime \u2013 Here\u2019s the 24/7 Number You Need",
     1
    ],
    [
     "21 Methods to Reach Live Support at Expedia",
     1
    ],
    [
     "Expedia Help Center: 15 Smart Ways to Solve Your Travel Problems",
     1
    ]
   ],
   "seconds": 88.0
  },
  "exclusion_rows": [
   {
    "what": "method entries removed as spam (name pattern)",
    "count": 12604,
    "why": "customer-service phrasing in the name, counted over the methods table and the paper catalogue's embedded method list; the pattern is public in the build script"
   },
   {
    "what": "duplicate leaderboard tables collapsed",
    "count": 10354,
    "why": "the same benchmark reached from several task nodes"
   },
   {
    "what": "method entries removed as spam (dummy-paper clusters)",
    "count": 6785,
    "why": "entries hung off five dummy paper records, counted over the methods table and the paper catalogue's embedded method list; see the attribution page"
   },
   {
    "what": "dataset rows removed as spam",
    "count": 2794,
    "why": "contact-centre spam in the dataset's name or full name; the pattern is public in the build script"
   },
   {
    "what": "papers excluded for having no title",
    "count": 636,
    "why": "a page needs a title"
   },
   {
    "what": "per-paper method tags dropped as spam",
    "count": 37766,
    "why": "the method-name pattern, applied to the tags the paper catalogue carries"
   },
   {
    "what": "dataset descriptions blanked as vandalised",
    "count": 201,
    "why": "row kept, text removed"
   },
   {
    "what": "method-collection descriptions blanked as vandalised",
    "count": 65,
    "why": "text removed"
   },
   {
    "what": "method descriptions blanked as vandalised",
    "count": 22,
    "why": "row kept, text removed"
   },
   {
    "what": "task descriptions blanked as vandalised",
    "count": 5,
    "why": "row kept, text removed"
   }
  ],
  "content_rule_rows": [
   {
    "what": "dataset rows withheld, all reasons",
    "count": 36
   },
   {
    "what": "dataset rows withheld: affiliate or product advertising",
    "count": 24
   },
   {
    "what": "dataset rows withheld: travel-booking or airline-support spam",
    "count": 7
   },
   {
    "what": "dataset rows withheld: gambling advertising",
    "count": 4
   },
   {
    "what": "dataset rows withheld: contact-centre spam",
    "count": 1
   },
   {
    "what": "method entries withheld, all reasons",
    "count": 31
   },
   {
    "what": "method entries withheld: contact-centre spam",
    "count": 10
   },
   {
    "what": "method entries withheld: travel-booking or airline-support spam",
    "count": 7
   },
   {
    "what": "method entries withheld: affiliate or product advertising",
    "count": 6
   },
   {
    "what": "method entries withheld: gambling advertising",
    "count": 3
   },
   {
    "what": "method entries withheld: pharmacy advertising",
    "count": 2
   },
   {
    "what": "method entries withheld: a test entry",
    "count": 2
   },
   {
    "what": "method entries withheld: name and text replaced by an advert (affiliate or product advertising)",
    "count": 1
   },
   {
    "what": "task entries withheld, all reasons",
    "count": 2
   },
   {
    "what": "task entries withheld: contact-centre spam",
    "count": 2
   },
   {
    "what": "leaderboard tables withheld, all reasons",
    "count": 27
   },
   {
    "what": "leaderboard tables withheld: its only row was a test entry",
    "count": 23
   },
   {
    "what": "leaderboard tables withheld: every row was withheld as spam",
    "count": 3
   },
   {
    "what": "leaderboard tables withheld: its task was withheld",
    "count": 1
   },
   {
    "what": "leaderboard rows withheld, all reasons",
    "count": 26
   },
   {
    "what": "leaderboard rows withheld: a single row credited to a dummy paper record",
    "count": 23
   },
   {
    "what": "leaderboard rows withheld: contact-centre spam",
    "count": 2
   },
   {
    "what": "leaderboard rows withheld: travel-booking or airline-support spam",
    "count": 1
   },
   {
    "what": "method entries: description blanked, row kept: one junk word as the whole text",
    "count": 2
   },
   {
    "what": "dataset rows: licence link blanked, row kept: it pointed at a spam or darknet-market host",
    "count": 1
   },
   {
    "what": "dataset rows: description blanked, row kept: darknet advertising",
    "count": 1
   },
   {
    "what": "dataset rows: licence name blanked, row kept: it repeated the spam name that replaced the dataset's own name in the same edit (the name was restored from the dataset's address; the licence link was removed with it)",
    "count": 1
   },
   {
    "what": "method entries: description blanked, row kept: a phone number",
    "count": 1
   },
   {
    "what": "task entries: description blanked, row kept: a phone number",
    "count": 1
   },
   {
    "what": "dataset names changed in the same edit as a spam description, restored from the dataset's address",
    "count": 1
   },
   {
    "what": "distinct method tag names on papers withheld as spam",
    "count": 255
   },
   {
    "what": "distinct task tag names on papers withheld as spam",
    "count": 1
   },
   {
    "what": "method \"see code\" links shown as plain text: not a code host",
    "count": 26
   },
   {
    "what": "method source links shown as plain text: not a paper host",
    "count": 12
   }
  ],
  "syntology_layer_rows": [
   {
    "what": "papers in Syntology's layer",
    "count": 54410,
    "unit": "papers"
   },
   {
    "what": "papers with harvested code samples",
    "count": 51887,
    "unit": "papers"
   },
   {
    "what": "papers newer than the archive with a repository link in the graph",
    "count": 9581,
    "unit": "papers"
   },
   {
    "what": "distinct code samples",
    "count": 294260,
    "unit": "samples"
   },
   {
    "what": "samples that ran (any status except unverified)",
    "count": 127244,
    "unit": "samples"
   },
   {
    "what": "ran, honoured the contract we drafted",
    "count": 8859,
    "unit": "samples"
   },
   {
    "what": "ran, violated the contract we drafted",
    "count": 3135,
    "unit": "samples"
   },
   {
    "what": "ran; our contract draft was wrong, not the code",
    "count": 24451,
    "unit": "samples"
   },
   {
    "what": "ran; our fixture could not drive it",
    "count": 9437,
    "unit": "samples"
   },
   {
    "what": "ran on a synthesized input",
    "count": 81362,
    "unit": "samples"
   },
   {
    "what": "unverified (harvested, no recorded run)",
    "count": 167016,
    "unit": "samples"
   },
   {
    "what": "samples with a behaviour fingerprint",
    "count": 37273,
    "unit": "samples"
   },
   {
    "what": "links whose sample has no function or class name",
    "count": 0,
    "unit": "links"
   },
   {
    "what": "paper\u2013sample links",
    "count": 361576,
    "unit": "links"
   },
   {
    "what": "links that are pointer only (Syntology does not serve that copy's text)",
    "count": 124983,
    "unit": "links"
   },
   {
    "what": "links where this paper's copy is the one Syntology first harvested",
    "count": 243453,
    "unit": "links"
   },
   {
    "what": "links whose repository and path the harvester recorded for this paper",
    "count": 100333,
    "unit": "links"
   },
   {
    "what": "links whose repository and path come from a harvest plan row for this paper",
    "count": 12948,
    "unit": "links"
   },
   {
    "what": "links to identical code first harvested for another paper (this paper's copy not located)",
    "count": 4842,
    "unit": "links"
   }
  ],
  "graph_layer_summary": {
   "distinct_samples_written": 294260,
   "link_basis:first_harvest_node": 243453,
   "link_basis:harvester_set": 100333,
   "link_basis:identical_code_first_harvested_elsewhere": 4842,
   "link_basis:plan_row": 12948,
   "links_written": 361576,
   "papers_total": 54410,
   "papers_with_harvested_samples": 51887,
   "papers_written": 54410,
   "post_freeze_repo_linked": 9581,
   "post_freeze_without_date": 9497,
   "samples_pointer_only": 124983,
   "seconds": 315.7,
   "status_ran": 149488,
   "status_ran_honours": 14024,
   "status_ran_violates": 6490,
   "status_unverified": 191574
  },
  "post_freeze_pages": {
   "total": 9581,
   "with_arxiv_metadata": 9280,
   "without_arxiv_metadata": 301,
   "without_repository_for_attached_rows": 0
  },
  "extracted_results": {
   "present": false,
   "source": "lb_extract (leaderboard-shaped extractor), verified",
   "file": "cwp/data/lb_rows_run_verified.jsonl",
   "file_sha256_as_read": null,
   "file_bytes_as_read": 0,
   "file_lines": {},
   "papers_in_file": 0,
   "status": {},
   "boards_proposed": 0,
   "refused_by_rule": {},
   "refused_by_rule_cells": 0,
   "verifier": {},
   "declined_by_site": {},
   "entries_accepted": 0,
   "cells_accepted": 0,
   "boards_with_entries": 0,
   "papers_with_entries": 0,
   "configuration_not_shown": {},
   "measurement": {
    "test_papers": 883,
    "papers_with_output": 881,
    "papers_with_output_from": null,
    "accepted_entries": 110,
    "judged_true": 108,
    "judged": 110,
    "wilson95_lower": 0.9361,
    "passed": true,
    "frozen_commit": "0e3de0df94",
    "result_commit": null,
    "measured_on": "2026-09-24"
   },
   "measurement_sha256_as_read": "f33f040865aa9b360a85c26e097b3f2d66cb34149574b9366039d37791554b75"
  },
  "after_reader_reports": {
   "removed": [],
   "by_kind": {
    "code_links_add": {},
    "code_links_remove": {},
    "archive_values": {},
    "titles": {},
    "suppress_syntology_rows": {},
    "suppress_samples": {},
    "notes": {}
   }
  }
 },
 "seconds": 444.7,
 "_provenance": {
  "schema": "syntology-provenance/1",
  "created_at": "2026-09-24T21:36:54+00:00",
  "generator": {
   "script": "cwp/build.py",
   "argv": [
    "--extracted"
   ],
   "git_commit": "8b66c1209bc7ffed9fb99e14a606a289c6b4dd6c",
   "git_dirty": true
  },
  "inputs": [
   {
    "path": "/Users/qed/Documents/syntology-final/cwp/data/catalog.sqlite",
    "sha256": "d43600e8e2de469295a395c7a2f349447b0850539effc2cd8932d430221af765",
    "bytes": 1809960960,
    "mtime": "2026-09-24T00:38:54+00:00"
   },
   {
    "path": "/Users/qed/Documents/syntology-final/cwp/data/graph_layer.jsonl",
    "sha256": "83751dfd6b2867f00fce8aad05c395c6b7b0157d062e384fdc62872d7a48c69e",
    "bytes": 190593408,
    "mtime": "2026-09-24T18:20:29+00:00"
   }
  ],
  "params": {
   "limit": 0,
   "only": "",
   "dist": "/Users/qed/Documents/syntology-final/cwp/dist",
   "data": "/Users/qed/Documents/syntology-final/cwp/data",
   "extracted": true,
   "participation": "auto",
   "atlas": "auto"
  }
 }
}