Append validated Tasksource decisions: sentfin, wands, scirepeval/search

README.md

@@ -80,16 +80,16 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 1531642708
-    num_examples: 1031657
+    num_bytes: 1536987728
+    num_examples: 1034657
   - name: validation
-    num_bytes: 31014339
-    num_examples: 15550
+    num_bytes: 31520421
+    num_examples: 15850
   - name: test
-    num_bytes: 31406839
-    num_examples: 16247
-  download_size: 576098327
-  dataset_size: 1594063886
+    num_bytes: 31713095
+    num_examples: 16447
+  download_size: 578643140
+  dataset_size: 1600221244
 - config_name: filtered-full
   features:
   - name: state
@@ -156,16 +156,16 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 3317682686
-    num_examples: 2529342
+    num_bytes: 3323027706
+    num_examples: 2532342
   - name: validation
-    num_bytes: 31014339
-    num_examples: 15550
+    num_bytes: 31520421
+    num_examples: 15850
   - name: test
-    num_bytes: 31406839
-    num_examples: 16247
-  download_size: 1252967330
-  dataset_size: 3380103864
+    num_bytes: 31713095
+    num_examples: 16447
+  download_size: 1255512143
+  dataset_size: 3386261222
 - config_name: vision
   features:
   - name: images
@@ -463,3 +463,6 @@ are unchanged. Current counts are in the config metadata. See
 ## HTML/browser text additions
 
 WebLINX action/DOM-element and WebSRC yes/no/element tasks retain their source partitions. WebLINX named test subsets are combined into the hosted test split, with native split names in row provenance. `mind2web/action` and `mind2web/dom-element` use only the original public training trajectories. Tasksource-derived train/validation/test holdouts keep whole trajectories and identical DOM snapshots together (80/10/10 deterministic group hashing). These are not the official benchmark test partitions. The same partition is used for both annotations. Original annotation/action IDs and page hashes are recorded in [row provenance](additions/167c5ad6fb79/rows.json). Only past actions enter context; four-way choices use native DOM candidates. Direct decisions are appended to default/full; existing shards, vision and filtered-full are unchanged. Alignment with the Decision Index private evaluation rows has not been independently verified.
+SEntFiN entity sentiment, WANDS ordered relevance grades and SciRepEval Search click-derived scores are added as direct decisions to default/full. Headline/query groups stay within one split. SEntFiN and WANDS use derived 80/10/10 holdouts; Search uses native train/validation with overlapping queries removed and no benchmark evaluation data. Search numeric scores are interpolated over 0,2,...,14 anchors, preserving their expectation; they are not human mean relevance ratings. RELISH and SciNUP are excluded as evaluation-only. Source code/provenance retain original score semantics and license evidence; Search text licensing remains unspecified.
+
+Source-row metadata: additions/2d27d40ae2fe/rows.json.

additions/2d27d40ae2fe/manifest.json

@@ -0,0 +1,206 @@
+{
+  "parent_revision": "477db7523772a988510c336c68cb623bfd989755",
+  "tasks": [
+    "sentfin",
+    "wands",
+    "scirepeval/search"
+  ],
+  "configs": [
+    "default",
+    "full"
+  ],
+  "max_rows": 1000,
+  "max_rows_eval": 100,
+  "splits": {
+    "train": {
+      "rows": 3000,
+      "sources": {
+        "sentfin": 1000,
+        "wands": 1000,
+        "scirepeval/search": 1000
+      },
+      "num_bytes": 5345020,
+      "file_bytes": 2207535,
+      "sha256": "349de62925013eb3de0da65881479d033ba2007b7fc95c4fcf467a30d9e1caf1"
+    },
+    "validation": {
+      "rows": 300,
+      "sources": {
+        "sentfin": 100,
+        "wands": 100,
+        "scirepeval/search": 100
+      },
+      "num_bytes": 506082,
+      "file_bytes": 218051,
+      "sha256": "7be7c7a5439ec6335241c59574e9767b045f93242ff6c1a5075fb29b29544fcd"
+    },
+    "test": {
+      "rows": 200,
+      "sources": {
+        "sentfin": 100,
+        "wands": 100
+      },
+      "num_bytes": 306256,
+      "file_bytes": 119227,
+      "sha256": "1bcf4e2366f95ab08b9b57c97fb572c5b78ff326eb841c638298dc81c4a0519f"
+    }
+  },
+  "provenance": {
+    "sentfin": {
+      "dataset": "tasksource/sentfin",
+      "revision": "dbd1d5665ff8d64254459a89a3a542c7728a3268"
+    },
+    "wands": {
+      "dataset": "tasksource/wands",
+      "revision": "96a856e9d45eb0e5847a07a026d07f2b13092733",
+      "originals": [
+        "napsternxg/wands"
+      ]
+    },
+    "scirepeval/search": {
+      "dataset": "tasksource/scirepeval-search",
+      "revision": "7ff532ab52c10274a818d812c8d47f63b3af7af4",
+      "originals": [
+        "allenai/scirepeval"
+      ]
+    }
+  },
+  "excluded": [],
+  "max_request_bytes": 131072,
+  "licenses": {
+    "sentfin": {
+      "license": "mit",
+      "license_use": "commercial",
+      "card_licenses": {
+        "tasksource/sentfin": [
+          "mit"
+        ]
+      }
+    },
+    "wands": {
+      "license": "mit",
+      "license_use": "commercial",
+      "card_licenses": {
+        "tasksource/wands": [
+          "mit"
+        ],
+        "napsternxg/wands": [
+          "mit"
+        ]
+      }
+    },
+    "scirepeval/search": {
+      "license": "other",
+      "license_use": "unspecified",
+      "card_licenses": {
+        "tasksource/scirepeval-search": [
+          "other"
+        ]
+      }
+    }
+  },
+  "conversion_sha256": "729160ad55fbd279ea71c35c5e9f1923e12e1719c1a007a8b33f1fef138f63ac",
+  "row_metadata": "additions/2d27d40ae2fe/rows.json",
+  "preserved_parquets": {
+    "data/test-00000-of-00001.parquet": "3db460225ea8c12502ae054af17da017069021f6",
+    "data/test-tasksource-addition-167c5ad6fb79.parquet": "5bda26554a6878ec629641bca4281bd091bbd09c",
+    "data/test-webinstruct-05e8b528d68c.parquet": "6ad9c90ee91f10605db36bf9973e753c7ac222d3",
+    "data/train-00000-of-00005.parquet": "a5ea112f200482f7d1a0fa09d0fa5c49723e91a1",
+    "data/train-00001-of-00005.parquet": "d875416ece2d448ee66af7d4d8c3bf6083c904b5",
+    "data/train-00002-of-00005.parquet": "4af8423c60163c1c6e85ab3bb9757124306e9c0f",
+    "data/train-00003-of-00005.parquet": "7cc5fcdb9c58fa7ec7fd8f8c565ec53bbdce9df1",
+    "data/train-00004-of-00005.parquet": "7a56772d889047eedd85bf5ef99aa9942dd875b5",
+    "data/train-tasksource-addition-167c5ad6fb79.parquet": "23740a966e0779b2f42517bdc59434979d69217e",
+    "data/train-webinstruct-05e8b528d68c.parquet": "0ca8787b5c6919fbd6c13332361fcdd94b90b005",
+    "data/validation-00000-of-00001.parquet": "f1f31f53bd083da0a5323223874c09583ecde6aa",
+    "data/validation-tasksource-addition-167c5ad6fb79.parquet": "3a7921a10f3cfb93459f29fd93ee56806b734726",
+    "filtered-full/train-00000-of-00013.parquet": "039fc948f281e3c8148fe004e681d735591cdea3",
+    "filtered-full/train-00001-of-00013.parquet": "f88b8e4697d520ca45c697326fab58ccbac540bf",
+    "filtered-full/train-00002-of-00013.parquet": "f5d9d7510ebcfdb0776cff688b7397326032a0c0",
+    "filtered-full/train-00003-of-00013.parquet": "4305c9327bc3a9fd835d889b453f0d0cde9869d5",
+    "filtered-full/train-00004-of-00013.parquet": "56a723d674dbdbc9a7550bfa4dfe27b3b15afe8b",
+    "filtered-full/train-00005-of-00013.parquet": "09360c342fd51e8b9cd5c3b46a0d80a9fcada073",
+    "filtered-full/train-00006-of-00013.parquet": "ac918bbe575e979b227a0edd72a0a832f923bec6",
+    "filtered-full/train-00007-of-00013.parquet": "da417ceefaf0197cf9da5f0bb5995890e00540c9",
+    "filtered-full/train-00008-of-00013.parquet": "e85c642b8494b662cbfb00a1a412d145e148524e",
+    "filtered-full/train-00009-of-00013.parquet": "1db9e70a1d778f3dbe8bd97b7470f3a864bfc79c",
+    "filtered-full/train-00010-of-00013.parquet": "bbb36286f0c84a763f90b00ac95d35d43a69ddf1",
+    "filtered-full/train-00011-of-00013.parquet": "4de7289f96ce6d8161c20b3bad4263b0d915f1e4",
+    "filtered-full/train-00012-of-00013.parquet": "cbcd789867f6c9f7b8a5df79cdcaac5a8d173736",
+    "full/test-00000-of-00001.parquet": "3db460225ea8c12502ae054af17da017069021f6",
+    "full/test-tasksource-addition-167c5ad6fb79.parquet": "5bda26554a6878ec629641bca4281bd091bbd09c",
+    "full/test-webinstruct-05e8b528d68c.parquet": "6ad9c90ee91f10605db36bf9973e753c7ac222d3",
+    "full/train-00000-of-00013.parquet": "b993ab1d3af81c78817f5a8a07ff5f00854cc6d9",
+    "full/train-00001-of-00013.parquet": "b3e533e478154b334ac57f61efed3d008074c76a",
+    "full/train-00002-of-00013.parquet": "21d0d533b037e85055f8ff777c05631682de8f0d",
+    "full/train-00003-of-00013.parquet": "67b1395b90e951196d4a97ec9a5bddc58769edfe",
+    "full/train-00004-of-00013.parquet": "f772ec552e89186430b329b62ac822f3b12377ed",
+    "full/train-00005-of-00013.parquet": "17568b8d8ddc14d4618a49c1cbee15dc4100e297",
+    "full/train-00006-of-00013.parquet": "d080d2007a13c164ff404826c5ddf2841edbc616",
+    "full/train-00007-of-00013.parquet": "b2937dc5b6f536338acade353d3b929d6645b1ee",
+    "full/train-00008-of-00013.parquet": "f7ecfa3628e6ac8624d038c738f14901625023f9",
+    "full/train-00009-of-00013.parquet": "c0012cd9c544647ee79cb85acda83068e8e719a6",
+    "full/train-00010-of-00013.parquet": "f4d78167ca607923cea1f05a9589091a040ae906",
+    "full/train-00011-of-00013.parquet": "51935b420cca5ebc1456ce82235df721d4357dad",
+    "full/train-00012-of-00013.parquet": "857f4c5cd01fb94a24ed594cf12bc6143801f146",
+    "full/train-tasksource-addition-167c5ad6fb79.parquet": "23740a966e0779b2f42517bdc59434979d69217e",
+    "full/train-webinstruct-05e8b528d68c.parquet": "0ca8787b5c6919fbd6c13332361fcdd94b90b005",
+    "full/validation-00000-of-00001.parquet": "f1f31f53bd083da0a5323223874c09583ecde6aa",
+    "full/validation-tasksource-addition-167c5ad6fb79.parquet": "3a7921a10f3cfb93459f29fd93ee56806b734726",
+    "quality/filtered-full-v1/candidates.parquet": "8b201f9bcdf4857d202ac540476e148dd02d9d69",
+    "quality/filtered-full-v1/excluded.parquet": "19c6ac1685ddc4c7228d13053dc8efeef6e3cc5e",
+    "vision/test-00000-of-00001.parquet": "f1a017f0f97d4d14c6c115eb296f6bcde5f29bc8",
+    "vision/test-addition-20261009-00000-of-00001.parquet": "ff9f1158392db36149122653dadf76abfac333a1",
+    "vision/train-00000-of-00036.parquet": "046dc0d2dbc23765e598fc4dfbe7bbbdd988a641",
+    "vision/train-00001-of-00036.parquet": "d471d0701b949ff85b487f6f551265b7d64afa48",
+    "vision/train-00002-of-00036.parquet": "df10056fde1d61165ca526c490517bf17f330deb",
+    "vision/train-00003-of-00036.parquet": "10717e54f825fa3dea50b56f690b3498436b4195",
+    "vision/train-00004-of-00036.parquet": "e7a052aa625aaf204d13fffe991022f0c4583e5b",
+    "vision/train-00005-of-00036.parquet": "5c5e2ae087bf1998ea7d2492f3ef0b284f95db1c",
+    "vision/train-00006-of-00036.parquet": "c2bef19550d56e4c1e34ba30bee40063ede42e01",
+    "vision/train-00007-of-00036.parquet": "3433079b9d8ebc06e76257d7f0d53633f09fd995",
+    "vision/train-00008-of-00036.parquet": "e2594e38e8335206d446864784a5fe5c8ad4d3fe",
+    "vision/train-00009-of-00036.parquet": "ea169b77941607243259a86557937e511e8e458e",
+    "vision/train-00010-of-00036.parquet": "922d16d0c673a4435bba94fb80045f26918758cd",
+    "vision/train-00011-of-00036.parquet": "2547f0d11f42b7a2244375c295cd7e9688e0097a",
+    "vision/train-00012-of-00036.parquet": "8142b2b6ebae02783c925f83f7a6954d557ce685",
+    "vision/train-00013-of-00036.parquet": "2f5e11e51d0d6a3798575d16aadbf8e1bf28424f",
+    "vision/train-00014-of-00036.parquet": "c59050c2c20c1f4086f68acc35cf7dafa6bfe1e9",
+    "vision/train-00015-of-00036.parquet": "eec5ad6feb41c31fa6e1971f1215de8c7098aec9",
+    "vision/train-00016-of-00036.parquet": "5d2dba81e43e89b79821701d2355819e5716146d",
+    "vision/train-00017-of-00036.parquet": "7075219f0d7fc3d53ff89fb9ef0f877a70a79b02",
+    "vision/train-00018-of-00036.parquet": "8212aa0fd96af09a231662f2003c3115f05642ce",
+    "vision/train-00019-of-00036.parquet": "1e9b1d605074d1cdfb0b3becf0a68677a97fbba8",
+    "vision/train-00020-of-00036.parquet": "42cd84c93abdc793cc2fa146cfe330d00c52933f",
+    "vision/train-00021-of-00036.parquet": "38a617eabd8f53dd93577aadc5a64c06481c57d1",
+    "vision/train-00022-of-00036.parquet": "9d3463391757063003d37bfa504b3f1729b6dfe4",
+    "vision/train-00023-of-00036.parquet": "3217c4ca2d7d280bf8d0f7cc782f00e3f5354640",
+    "vision/train-00024-of-00036.parquet": "06d6d89788b3d94e3071dca6bd360724aab6a4eb",
+    "vision/train-00025-of-00036.parquet": "e00e2cdaa5fde114a66edef069fb2e769723f12a",
+    "vision/train-00026-of-00036.parquet": "075e6dd30dec0c59abb20d135d021d200bec2af4",
+    "vision/train-00027-of-00036.parquet": "8ed3500cebc811e4abe52b7859428c3a4cd68c2a",
+    "vision/train-00028-of-00036.parquet": "8c31114b2d8869e9b861f1b8f88bc1012a7cc706",
+    "vision/train-00029-of-00036.parquet": "f780cfb0a3b8f618699d3870a8a4cac53c1b2443",
+    "vision/train-00030-of-00036.parquet": "6afa24ed5004f14624ac25a6d0271a49c42f0328",
+    "vision/train-00031-of-00036.parquet": "94ca2674a0bb8fe778085bb2ff246e82341b2762",
+    "vision/train-00032-of-00036.parquet": "e5e6c49dad8b8aba932b51085402252f271bb724",
+    "vision/train-00033-of-00036.parquet": "3024eb6074cf43e9e9b26b1c694275fc358bcbbc",
+    "vision/train-00034-of-00036.parquet": "0a4a67293feef6d9c35567945b69db9168ec424e",
+    "vision/train-00035-of-00036.parquet": "2d67eff8bcd5d6bb1a4e846771db5fea86460fb2",
+    "vision/train-addition-20261009-00000-of-00010.parquet": "31c6632e1fee045cb3f445a1fe3398316b5d8a2b",
+    "vision/train-addition-20261009-00001-of-00010.parquet": "4616e739fcb0f0aac41afa75723cef2c48a963a8",
+    "vision/train-addition-20261009-00002-of-00010.parquet": "56558d76e3577319808282354cbc74407f2957ca",
+    "vision/train-addition-20261009-00003-of-00010.parquet": "506a53fed531e2370d787d9aa630977868a20405",
+    "vision/train-addition-20261009-00004-of-00010.parquet": "b9b646aaf42064ef674c69025e645ccd4425105d",
+    "vision/train-addition-20261009-00005-of-00010.parquet": "335371343dc775a49ca3873e904ed4640ac62e26",
+    "vision/train-addition-20261009-00006-of-00010.parquet": "53e15fb9a7ee15457bb0132777ffff420bbd5f4b",
+    "vision/train-addition-20261009-00007-of-00010.parquet": "742d65c41133deec13b0fcc4e5e6052a375cb9f6",
+    "vision/train-addition-20261009-00008-of-00010.parquet": "5e150c32add03b1e8edc0fd3b693a8b0485aefce",
+    "vision/train-addition-20261009-00009-of-00010.parquet": "0a135d907174ad5fc7aefffce16e6e69d94fe6f1",
+    "vision/validation-00000-of-00001.parquet": "d90a3706877935b558e1257b128dc3526ec88f97",
+    "vision/validation-addition-20261009-00000-of-00001.parquet": "182b85299abf9a74728c600dd2f576233cc06e15"
+  },
+  "git_revision": "1597ac91c250fd9a3f1fc1fa28bba5ec510d376b",
+  "task_definitions_sha256": "7d14c44b71286c4247bb1e02b535a1e098be1d4031cfff50f8a0c6ece19d180d",
+  "quality_audit": "quality-audit.json"
+}

additions/2d27d40ae2fe/quality-audit.json

@@ -0,0 +1,286 @@
+{
+  "scope": "Three staged training examples per task; source-label alignment on every emitted row; full mirror group-disjointness.",
+  "source_label_alignment_rows": {
+    "sentfin": 1200,
+    "wands": 1200,
+    "scirepeval/search": 1100
+  },
+  "split_audit": {
+    "sentfin": {
+      "group_counts": {
+        "test": 1063,
+        "train": 8521,
+        "validation": 1095
+      },
+      "label_counts": {
+        "test": {
+          "negative": 370,
+          "neutral": 549,
+          "positive": 498
+        },
+        "train": {
+          "neutral": 4339,
+          "positive": 4014,
+          "negative": 3059
+        },
+        "validation": {
+          "positive": 542,
+          "neutral": 578,
+          "negative": 369
+        }
+      },
+      "cross_split_group_overlap": 0
+    },
+    "wands": {
+      "group_counts": {
+        "test": 50,
+        "train": 375,
+        "validation": 55
+      },
+      "label_counts": {
+        "test": {
+          "1": 18501,
+          "0": 7261,
+          "2": 2843
+        },
+        "train": {
+          "1": 113033,
+          "2": 20289,
+          "0": 49731
+        },
+        "validation": {
+          "1": 15099,
+          "2": 2482,
+          "0": 4209
+        }
+      },
+      "cross_split_group_overlap": 0
+    },
+    "scirepeval-search": {
+      "group_counts": {
+        "train": 4967,
+        "validation": 494
+      },
+      "cross_split_group_overlap": 0
+    }
+  },
+  "training_label_counts": {
+    "sentfin": {
+      "neutral": 403,
+      "negative": 244,
+      "positive": 353
+    },
+    "wands": {
+      "1": 621,
+      "2": 100,
+      "0": 279
+    },
+    "scirepeval/search": {
+      "0.0": 887,
+      "1.0": 108,
+      "2.0": 4,
+      "3.0": 1
+    }
+  },
+  "examples": {
+    "sentfin": [
+      {
+        "id": "sentfin:train:5058:Reliance Infrastructure:sentfin-a665621803",
+        "state": "text_A: Brokers' call on  Exide Industries, Bharti Airtel, Reliance Infrastructure\ntext_B: Reliance Infrastructure",
+        "question": "What sentiment does headline text_A express toward entity text_B?",
+        "kind": "choice",
+        "options": [
+          "negative",
+          "neutral",
+          "positive"
+        ],
+        "target": [
+          0.0,
+          1.0,
+          0.0
+        ],
+        "native_label": "neutral"
+      },
+      {
+        "id": "sentfin:train:6870:NTPC:sentfin-a665621803",
+        "state": "text_A: 191-192 is where price of NTPC to come in next few trading sessions: Sandeep Wagle\ntext_B: NTPC",
+        "question": "What sentiment does headline text_A express toward entity text_B?",
+        "kind": "score",
+        "options": [
+          "negative",
+          "neutral",
+          "positive"
+        ],
+        "target": [
+          0.0,
+          1.0,
+          0.0
+        ],
+        "native_label": "neutral"
+      },
+      {
+        "id": "sentfin:train:8685:two-wheeler:sentfin-a665621803",
+        "state": "text_A: Cautious on two-wheeler space: David Pezarkar\ntext_B: two-wheeler",
+        "question": "What sentiment does headline text_A express toward entity text_B?",
+        "kind": "score",
+        "options": [
+          "negative",
+          "neutral",
+          "positive"
+        ],
+        "target": [
+          1.0,
+          0.0,
+          0.0
+        ],
+        "native_label": "negative"
+      }
+    ],
+    "wands": [
+      {
+        "id": "wands:train:146017:wands-a340aa1a78",
+        "state": "text_A: tressler rug\ntext_B: product_name: geometric machine made power loom wool and polyester red/green/black area rug\nproduct_class: Area Rugs\ncategory hierarchy: Rugs / Area Rugs\nproduct_features: overallproductweight:3.2|material : wool|supplierintendedandapproveduse : non residential use|pileheight:0.35|overallproductweight:12.8|overallproductweight:6|productresistances : fade resistant|construction : machine made|overallproductweight:10|overallwidth-sidetoside:60|technique : power loom|overallwidth-sidetoside:48|overalllength-endtoend:60|productcare : vacuum with no beater bar/rotating brush|backingmaterialdetails : cotton|overalllength-endtoend:36|overallwidth-sidetoside:36|overalllength-endtoend:48|overallwidth-sidetoside : |material : polyester|overallproductweight:2.4|materialdetails : wool and polyester blend|countryoforigin : turkey|overalllength-endtoend : |overallproductweight:4|overallwidth-sidetoside:24|primarycolor : red/green/black|holidayoccasion : no holiday|location : indoor use only|one-of-a-kind : no|satraapproved : no|commercialwarranty : no|usdaorganicnopcertified : no|ista3aor6acertified : no|fringetassel : no|gots50globalorganictextilestandardcompliant : no|reversible : no|rugsampleavailable : no| : no|gols30globalorganiclatexstandardcertified : no|rugsets : no|high-low : no|licensedproduct : no|stainresistant : yes|cpsc-16cfr1630compliant : yes|sor2016-176compliant : yes|uniformpackagingandlabelingregulationsuplrcompliant : yes|generalcertificateofconformitygcc : yes|backingmaterial : yes|cpsc-16cfr1631compliant : yes|rugpadrecommended : yes|floorheatingsafe : yes|purposefuldistressingtype : no distressing|supplierintendedandapproveduse : residential use",
+        "question": "How relevant is product text_B to search query text_A, from irrelevant to partial to exact match?",
+        "kind": "score",
+        "options": [
+          "irrelevant",
+          "partial match",
+          "exact match"
+        ],
+        "target": [
+          0.0,
+          1.0,
+          0.0
+        ],
+        "native_label": 1
+      },
+      {
+        "id": "wands:train:152499:wands-a340aa1a78",
+        "state": "text_A: togo chair\ntext_B: product_name: guthrie 25.2 '' wide side chair\ncategory hierarchy: Furniture / Living Room Furniture / Chairs & Seating / Accent Chairs\nproduct_description: this slipper accent chair ( sofa ) has a tilted , reclining position that 's perfect for placing in a cozy reading or for adding extra seating in the living room . it 's built on a powder-coated metal sled base with a black finish for an industrial-inspired look . the armless back and seat are filled with foam for tons of support during movie night . there are two upholstery materials - microfiber or polyurethane for you to choose to suit your style . plus , the stitched edges on this chair adds an extra tactile touch .\nproduct_features: seatwidth-sidetoside:25.2|cleaningcode : x : clean by lightly brushing and/or vacuuming to remove dust|seatdepth-fronttoback:21.38|productcare : wipe clean with a microfiber cloth|countryoforigin-additionaldetails : made in china|overallheight-toptobottom:31.1|warrantylength:1 year|levelofassembly : partial assembly|overalldepth-fronttoback:33.8|overallwidth-sidetoside:25.2|estimatedtimetosetup:20|waterrepellant : no resiliency|design : side chair|warrantydetails:1 year limited warranty|cushionconstruction : foam|additionaltoolsrequirednotincluded : all tools included|upholsterycolor : brown|legcolor : black|overallproductweight:19.1|upholsterymaterial : polyurethane|seatheight-floortoseat:16.1|backtype : tight back|backfillmaterial : foam|minimumdoorwidth-sidetoside:34|legmaterialdetails : steel|pattern : solid color|upholsterymaterialdetails : microfiber or polyurethane|upholsterymaterial : microfiber/microsuede|framematerialdetails : steel|seatfillmaterial : foam|weightcapacity:264|backheight-seattotopofback:15|numberofchairsincluded:1|suggestednumberofpeopleforassemblyinstall:1|tosspillowsincluded : no|scscertified : no|ansibifmax54loungepublicseating : no|slipcovered : no|compositewoodproductcwp : no|nailheadtrim : no|calgreencompliant : no|ista3aor6acertified : no|removablecushions : no|cushionticking : no|armsincluded : no|canadaproductrestriction : no|ista1acertified : no|nfpacompliant : no|bhftiupholsteredfurnitureflammabilitycompliant : no|taacompliant : no| : no|iso14000certified : no|doublerubtested : no|firerated : no|fireresistant : no|iso14001certified : no|commercialwarranty : no|containsflameretardantmaterials : no|gsaapproved : no|tuftedcushions : no|mattressincluded : no|swivel : no|iso9000certified : no|organic : no|generalcertificateofconformitygcc : no|uniformpackagingandlabelingregulationsuplrcompliant : no|ottomanincluded : no|iso9001certified : no|caltb116compliant : yes|productwarranty : yes|caltb117-2013compliant : yes|eligibleforhardwarepacks : yes|adultassemblyrequired : yes|weightcapacitygreaterthan250lbs : yes|upholstered : yes|countryoforigin : china|framematerial : metal|legmaterial : metal|fullorlimitedwarranty : limited|purposefuldistressingtype : no distressing|supplierintendedandapproveduse : residential use",
+        "question": "How relevant is product text_B to search query text_A, from irrelevant to partial to exact match?",
+        "kind": "score",
+        "options": [
+          "irrelevant",
+          "partial match",
+          "exact match"
+        ],
+        "target": [
+          0.0,
+          1.0,
+          0.0
+        ],
+        "native_label": 1
+      },
+      {
+        "id": "wands:train:18637:wands-a340aa1a78",
+        "state": "text_A: block coffee table\ntext_B: product_name: ranchester coffee table\nproduct_class: Coffee & Cocktail Tables\ncategory hierarchy: Furniture / Living Room Furniture / Coffee Tables & End Tables / Coffee Tables / Square Coffee Tables\nproduct_description: crafted from concrete in a neutral patina finish , this understated coffee table anchors your living room layout in industrial style . measuring 17 '' h x 43.25 '' l x 43.25 '' w , this sizable square piece provides plenty of space to set down snacks , magazines , and more . a water-based acrylic surface sealer outfits this design to prevent water and stains from sinking in . since this hefty table weighs 189 lbs. , we recommend you grab a friend for easier assembly .\nproduct_features: overallproductweight:189|basematerialdetails : concrete|dsprimaryproductstyle : industrial|topcolor : ivory|countryoforigin : viet nam|topshape : square|style : industrial|basematerial : concrete|producttype : coffee table|dssecondaryproductstyle : contemporary industrial|topmaterialdetails : fiber reinforced natural concrete|basetype : pedestal|estimatedtimetoassembleinternal:100|overalllength-endtoend:43.25|overalldepth-fronttoback:43.25|topcolor : dark grey|overallheight-toptobottom:17|topmaterial : concrete|numberoftablesincluded:1|cabinets : no|storageincluded : no|wheelsincluded : no|upholstered : no|openbase : no|drawersincluded : no|plug-in : no|drawers : no|shelving : no|stool : no|mirrored : no|stoolsincluded : no|folding : no| : no|cabinetsincluded : no|shelvesincluded : no|extendable : no|traytop : no|compositewoodproductcwp : no|lifttop : no|adultassemblyrequired : yes|purposefuldistressingtype : no distressing|naturalvariationtype : no natural variation",
+        "question": "How relevant is product text_B to search query text_A, from irrelevant to partial to exact match?",
+        "kind": "score",
+        "options": [
+          "irrelevant",
+          "partial match",
+          "exact match"
+        ],
+        "target": [
+          0.0,
+          1.0,
+          0.0
+        ],
+        "native_label": 1
+      }
+    ],
+    "scirepeval/search": [
+      {
+        "id": "scirepeval:train:19d73ee1bf4748c195b4cfb72c954610:158656248:scirepeval-search-c3128f518a",
+        "state": "text_A: New Pirates, Old Solutions: An 18th Century Approach to Cybercrime\ntext_B: Cybercrime: An Annotated Bibliography of Select Foreign Language Academic Literature\nCyberspace and Cybersecurity/Targets and vulnerabilities; Cyberspace and Cybersecurity/Policy and strategy",
+        "question": "Predict the click-derived search score of paper text_B for query text_A. Numeric criteria are score anchors, not human rating categories.",
+        "kind": "score",
+        "options": [
+          "0",
+          "2",
+          "4",
+          "6",
+          "8",
+          "10",
+          "12",
+          "14"
+        ],
+        "target": [
+          1.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "native_label": 0.0
+      },
+      {
+        "id": "scirepeval:train:3327bdab2f831bdfb76bcd28efae600e:27112207:scirepeval-search-c3128f518a",
+        "state": "text_A: a retrospective study on using a novel single needle cone puncture\ntext_B: Bioengineered vascular access maintains structural integrity in response to arteriovenous flow and repeated needle puncture.\nOBJECTIVE Tissue engineered blood vessels (TEBV) have been proposed as an alternative to prosthetic grafts for dialysis access. However, arteriovenous (AV) grafts must withstand extreme flow rates and frequent needle trauma. In a proof of concept study, we sought to determine whether scaffold based TEBV could withstand the hemodynamic and mechanical challenges of chronic dialysis access. METHODS TEBV were constructed using decellularized arterial scaffolds seeded with autologous ovine endothelial cells (EC) derived from circulating endothelial progenitor cells (EPC) using a novel high affinity capture approach. Seeded scaffolds were preconditioned to arterial pressure and flow in a bioreactor for 2 weeks prior to implantation to create carotid artery to jugular vein AV grafts in each animal. TEBV were healed for 1 month before initiating percutaneous needle puncture 3 days/week. TEBV wall geometry and patency were monitored using duplex imaging and were either explanted for histologic analysis at 2 months (n 5) or followed for up to 6 months until venous outflow stenosis threatened AV graft patency (n 6) RESULTS Despite high flow, TEBV maintained stable geometry with only modest wall dilation (under 6% by 4 months after implantation. Needle access was well tolerated with a single puncture site complication, a small pseudoaneurysm, occurring in the late group. Time to hemostasis at puncture sites averaged 4 2 minutes. Histologic analysis at 2 months demonstrated repopulation of the outer TEBV wall by host cells and healing of needle punctures by cellular ingrowth and new matrix deposition along the tract. TEBV followed beyond 2 months showed stable wall geometry but, consistent with the primary mode of clinical AV graft failure, all TEBV eventually developed venous anastomotic stenosis (mean, 4.4 0.9 months; range, 3.3 5.6 months postimplantation; n 6) CONCLUSIONS This pilot study supports the concept of creating dialysis access from scaffold based autologous TEBV. Engineered AV grafts were created within a clinically relevant time frame and demonstrated stable wall geometry despite high flow and repeated puncture. Cellular ingrowth and puncture site healing may improve wall durability, but venous outflow stenosis remains the primary mode of TEBV graft failure in the ovine model.",
+        "question": "Predict the click-derived search score of paper text_B for query text_A. Numeric criteria are score anchors, not human rating categories.",
+        "kind": "score",
+        "options": [
+          "0",
+          "2",
+          "4",
+          "6",
+          "8",
+          "10",
+          "12",
+          "14"
+        ],
+        "target": [
+          1.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "native_label": 0.0
+      },
+      {
+        "id": "scirepeval:train:b0d3ce4c6ef3dfcd3971037b44bcbdae:215893739:scirepeval-search-c3128f518a",
+        "state": "text_A: Bayesian Modeling of Dynamic Scenes for Object Detection\ntext_B: Bayesian Modeling of Dynamic Scenes for Object Detection\nAccurate detection of moving objects is an important precursor to stable tracking or recognition. In this paper, we present an object detection scheme that has three innovations over existing appro.",
+        "question": "Predict the click-derived search score of paper text_B for query text_A. Numeric criteria are score anchors, not human rating categories.",
+        "kind": "score",
+        "options": [
+          "0",
+          "2",
+          "4",
+          "6",
+          "8",
+          "10",
+          "12",
+          "14"
+        ],
+        "target": [
+          1.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "native_label": 0.0
+      }
+    ]
+  },
+  "manual_findings": {
+    "sentfin": "Three staged examples retain explicit entity questions and plausible native neutral/negative labels. A separate pilot example, Brokers Call (Cipla), has positive sentiment without explicit directional evidence; original labels can be context-dependent. No relabeling performed.",
+    "wands": "Three staged partial-match judgments keep complete product evidence and correct ordinal ordering; brand-specific queries can need product knowledge. Native fields are verbose.",
+    "scirepeval_search": "Inputs preserve query and paper evidence. Zero means native click score, not explicit irrelevance: one zero-score pair has identical query/paper title. Source scores are weak behavioral supervision; not human ratings."
+  },
+  "excluded_sources": {
+    "RELISH": "Evaluation-only; no derived training data",
+    "SciNUP": "Evaluation collection; no native training partition"
+  },
+  "limitations": "Coarse semantic inspection, not exhaustive answer verification; Search licensing remains unspecified."
+}

data/test-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bcf4e2366f95ab08b9b57c97fb572c5b78ff326eb841c638298dc81c4a0519f
+size 119227

data/train-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:349de62925013eb3de0da65881479d033ba2007b7fc95c4fcf467a30d9e1caf1
+size 2207535

data/validation-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be7c7a5439ec6335241c59574e9767b045f93242ff6c1a5075fb29b29544fcd
+size 218051

full/test-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bcf4e2366f95ab08b9b57c97fb572c5b78ff326eb841c638298dc81c4a0519f
+size 119227

full/train-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:349de62925013eb3de0da65881479d033ba2007b7fc95c4fcf467a30d9e1caf1
+size 2207535

full/validation-tasksource-addition-2d27d40ae2fe.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be7c7a5439ec6335241c59574e9767b045f93242ff6c1a5075fb29b29544fcd
+size 218051

release-audit.json

@@ -8,7 +8,7 @@
   },
   "splits": {
     "train": {
-      "rows": 2529342,
+      "rows": 2532342,
       "sources": {
         "AES2-essay-scoring": 3080,
         "AdjectiveScaleProbe-nli": 5017,
@@ -676,15 +676,18 @@
         "weblinx/action": 1000,
         "weblinx/dom-element": 1000,
         "websrc/yesno": 1000,
-        "websrc/element": 1000
+        "websrc/element": 1000,
+        "sentfin": 1000,
+        "wands": 1000,
+        "scirepeval/search": 1000
       },
       "kinds": {
-        "score": 187374,
-        "choice": 1979119,
+        "score": 189858,
+        "choice": 1979635,
         "noul": 362849
       },
       "variants": {
-        "direct": 2183487,
+        "direct": 2186487,
         "label_verification": 81080,
         "instruction_paraphrase": 38875,
         "paired_text_format": 27344,
@@ -693,8 +696,8 @@
       },
       "license_use": {
         "non-commercial": 257645,
-        "unspecified": 837789,
-        "commercial": 1433908
+        "unspecified": 838789,
+        "commercial": 1435908
       },
       "families": {
         "AES2-essay-scoring": 3080,
@@ -1055,10 +1058,13 @@
         "webinstruct": 28440,
         "mind2web": 1596,
         "websrc": 2000,
-        "weblinx": 2000
+        "weblinx": 2000,
+        "sentfin": 1000,
+        "scirepeval": 1000,
+        "wands": 1000
       },
       "formats": {
-        "Classification": 1204678,
+        "Classification": 1207678,
         "MultipleChoice": 786480,
         "graded": 250000,
         "procedural": 249999,
@@ -1066,7 +1072,7 @@
       }
     },
     "validation": {
-      "rows": 15550,
+      "rows": 15850,
       "sources": {
         "HelpSteer3/edit_quality": 24,
         "HelpSteer3/preference": 25,
@@ -1718,15 +1724,18 @@
         "weblinx/action": 100,
         "weblinx/dom-element": 100,
         "websrc/yesno": 100,
-        "websrc/element": 100
+        "websrc/element": 100,
+        "sentfin": 100,
+        "wands": 100,
+        "scirepeval/search": 100
       },
       "kinds": {
-        "choice": 12478,
+        "choice": 12518,
         "noul": 2063,
-        "score": 1009
+        "score": 1269
       },
       "variants": {
-        "direct": 13627,
+        "direct": 13927,
         "criteria_permutation": 443,
         "label_verification": 455,
         "instruction_paraphrase": 235,
@@ -1734,9 +1743,9 @@
         "packed_derived": 616
       },
       "license_use": {
-        "commercial": 8922,
+        "commercial": 9122,
         "non-commercial": 1812,
-        "unspecified": 4816
+        "unspecified": 4916
       },
       "families": {
         "HelpSteer3": 77,
@@ -2095,10 +2104,13 @@
         "yelp_review_full": 28,
         "mind2web": 173,
         "websrc": 200,
-        "weblinx": 200
+        "weblinx": 200,
+        "sentfin": 100,
+        "scirepeval": 100,
+        "wands": 100
       },
       "formats": {
-        "Classification": 7305,
+        "Classification": 7605,
         "MultipleChoice": 4795,
         "graded": 1500,
         "procedural": 1500,
@@ -2106,7 +2118,7 @@
       }
     },
     "test": {
-      "rows": 16247,
+      "rows": 16447,
       "sources": {
         "HelpSteer3/edit_quality": 26,
         "HelpSteer3/preference": 23,
@@ -2757,15 +2769,17 @@
         "mind2web/action": 77,
         "mind2web/dom-element": 72,
         "weblinx/action": 500,
-        "weblinx/dom-element": 500
+        "weblinx/dom-element": 500,
+        "sentfin": 100,
+        "wands": 100
       },
       "kinds": {
-        "choice": 13147,
+        "choice": 13195,
         "noul": 2127,
-        "score": 973
+        "score": 1125
       },
       "variants": {
-        "direct": 14291,
+        "direct": 14491,
         "criteria_permutation": 492,
         "label_verification": 451,
         "instruction_paraphrase": 227,
@@ -2773,7 +2787,7 @@
         "paired_text_format": 134
       },
       "license_use": {
-        "commercial": 8932,
+        "commercial": 9132,
         "non-commercial": 2625,
         "unspecified": 4690
       },
@@ -3134,10 +3148,12 @@
         "yelp_review_full": 28,
         "webinstruct": 122,
         "mind2web": 149,
-        "weblinx": 1000
+        "weblinx": 1000,
+        "sentfin": 100,
+        "wands": 100
       },
       "formats": {
-        "Classification": 7643,
+        "Classification": 7843,
         "MultipleChoice": 5154,
         "graded": 1500,
         "procedural": 1500,
@@ -3147,16 +3163,16 @@
   },
   "mix": {
     "test": {
-      "rows": 16247,
-      "sources": 650
+      "rows": 16447,
+      "sources": 652
     },
     "train": {
-      "rows": 1031657,
-      "sources": 667
+      "rows": 1034657,
+      "sources": 670
     },
     "validation": {
-      "rows": 15550,
-      "sources": 651
+      "rows": 15850,
+      "sources": 654
     }
   },
   "incremental_additions": [
@@ -3184,6 +3200,19 @@
         "validation": 573,
         "test": 1149
       }
+    },
+    {
+      "tasks": [
+        "sentfin",
+        "wands",
+        "scirepeval/search"
+      ],
+      "manifest": "additions/2d27d40ae2fe/manifest.json",
+      "rows": {
+        "train": 3000,
+        "validation": 300,
+        "test": 200
+      }
     }
   ],
   "english_task_rebuild": {

sources.yaml

@@ -74,6 +74,7 @@ datasets:
 - allenai/scicite
 - allenai/scifact_entailment
 - allenai/sciq
+- allenai/scirepeval
 - allenai/scitail
 - allenai/social_i_qa
 - allenai/swag
@@ -177,6 +178,7 @@ datasets:
 - multilingual-discourse-hub/disrpt
 - mwong/fever-evidence-related
 - nala-cub/americas_nli
+- napsternxg/wands
 - nedjmaou/MLMA_hate_speech
 - neurae/dnd_style_intents
 - neuralchemy/Prompt-injection-dataset
@@ -344,11 +346,13 @@ datasets:
 - tasksource/scicite
 - tasksource/scifact_entailment
 - tasksource/scinli
+- tasksource/scirepeval-search
 - tasksource/scone
 - tasksource/scruples
 - tasksource/seahorse_summarization_evaluation
 - tasksource/semantic_fragments_nli
 - tasksource/sen-making
+- tasksource/sentfin
 - tasksource/sharc
 - tasksource/sherliic
 - tasksource/sick
@@ -372,6 +376,7 @@ datasets:
 - tasksource/twentyquestions
 - tasksource/tydi-as2-balanced
 - tasksource/universal-joy
+- tasksource/wands
 - tasksource/webinstruct
 - tasksource/websrc
 - tasksource/wice
@@ -9075,3 +9080,50 @@ sources:
       - cc-by-4.0
       X-LANCE/WebSRC_v1.0:
       - cc-by-4.0
+  sentfin:
+    dataset: tasksource/sentfin
+    revision: dbd1d5665ff8d64254459a89a3a542c7728a3268
+    rows:
+      train: 1000
+      validation: 100
+      test: 100
+    revisions:
+      tasksource/sentfin: dbd1d5665ff8d64254459a89a3a542c7728a3268
+    license: mit
+    license_use: commercial
+    card_licenses:
+      tasksource/sentfin:
+      - mit
+  wands:
+    dataset: tasksource/wands
+    revision: 96a856e9d45eb0e5847a07a026d07f2b13092733
+    originals:
+    - napsternxg/wands
+    rows:
+      train: 1000
+      validation: 100
+      test: 100
+    revisions:
+      tasksource/wands: 96a856e9d45eb0e5847a07a026d07f2b13092733
+    license: mit
+    license_use: commercial
+    card_licenses:
+      tasksource/wands:
+      - mit
+      napsternxg/wands:
+      - mit
+  scirepeval/search:
+    dataset: tasksource/scirepeval-search
+    revision: 7ff532ab52c10274a818d812c8d47f63b3af7af4
+    originals:
+    - allenai/scirepeval
+    rows:
+      train: 1000
+      validation: 100
+    revisions:
+      tasksource/scirepeval-search: 7ff532ab52c10274a818d812c8d47f63b3af7af4
+    license: other
+    license_use: unspecified
+    card_licenses:
+      tasksource/scirepeval-search:
+      - other