{
  "title": "CMDB-1500",
  "question_count": 1500,
  "source_or_subtask_count": 79,
  "source_definition": "79 source and subtask identifiers; multiple subtasks may share an upstream dataset.",
  "sources": [
    {
      "source": "AI2D",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "test": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/lmms-lab/ai2d",
      "description": "Science diagram multiple-choice questions."
    },
    {
      "source": "ALFRED-high-level-action",
      "selected_questions": 19,
      "sector": "embodied_actions",
      "original_splits": {
        "dev": 19
      },
      "kinds": {
        "choice": 19
      },
      "upstream_url": "https://github.com/askforalfred/alfred",
      "description": "Select the demonstrated next high-level action from a task and previous actions."
    },
    {
      "source": "ALFWorld-expert-action-full",
      "selected_questions": 18,
      "sector": "embodied_actions",
      "original_splits": {
        "dev": 18
      },
      "kinds": {
        "choice": 18
      },
      "upstream_url": "https://github.com/alfworld/alfworld",
      "description": "Select the demonstrated action from a task and textual observation."
    },
    {
      "source": "Aegis-annotation",
      "selected_questions": 44,
      "sector": "safety",
      "original_splits": {
        "test": 44
      },
      "kinds": {
        "choice": 44
      },
      "upstream_url": "https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-1.0",
      "description": "Content safety classification using human annotations."
    },
    {
      "source": "BFCL-v3",
      "selected_questions": 25,
      "sector": "tools",
      "original_splits": {
        "test": 10,
        "dev": 15
      },
      "kinds": {
        "choice": 25
      },
      "upstream_url": "https://huggingface.co/datasets/gorilla-llm/Berkeley-Function-Calling-Leaderboard",
      "description": "Judge whether a tool call is needed."
    },
    {
      "source": "ChineseHP",
      "selected_questions": 50,
      "sector": "chinese_speech_candidate_judgment",
      "original_splits": {
        "test": 50
      },
      "kinds": {
        "choice": 50
      },
      "upstream_url": "https://github.com/tzyll/ChineseHP",
      "description": "Select among AISHELL-4 N-best text hypotheses using reference-based character edit distance."
    },
    {
      "source": "DecisionBench/AGT-1",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which tool to call",
      "upstream_sources": [
        {
          "dataset_id": "bfcl-v3-live",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#bfcl-v3-live",
          "dataset_name": "BFCL v3 Live (Berkeley Function-Calling Leaderboard)"
        }
      ],
      "description": "Select a tool for a user request"
    },
    {
      "source": "DecisionBench/AGT-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Did the injection succeed?",
      "upstream_sources": [
        {
          "dataset_id": "agentdojo",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#agentdojo",
          "dataset_name": "AgentDojo recorded runs"
        }
      ],
      "description": "Judge whether a prompt injection hijacked an agent"
    },
    {
      "source": "DecisionBench/AGT-3",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Did the agent complete the task?",
      "upstream_sources": [
        {
          "dataset_id": "tau-bench",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#tau-bench",
          "dataset_name": "τ-bench historical trajectories"
        }
      ],
      "description": "Judge whether an agent completed the user task"
    },
    {
      "source": "DecisionBench/AGT-4",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Is the answer backed by sources?",
      "upstream_sources": [
        {
          "dataset_id": "hagrid",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#hagrid",
          "dataset_name": "HAGRID"
        }
      ],
      "description": "Verify whether cited evidence supports an answer"
    },
    {
      "source": "DecisionBench/AGT-5",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which reply is better?",
      "upstream_sources": [
        {
          "dataset_id": "mt-bench-human",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#mt-bench-human",
          "dataset_name": "MT-Bench human judgments"
        }
      ],
      "description": "Compare the quality of two responses"
    },
    {
      "source": "DecisionBench/COM-1",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Does this product match the query?",
      "upstream_sources": [
        {
          "dataset_id": "amazon-esci",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#amazon-esci",
          "dataset_name": "Amazon Shopping Queries Dataset (ESCI)"
        }
      ],
      "description": "Assess product relevance to a shopping query"
    },
    {
      "source": "DecisionBench/COM-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What type of product is this?",
      "upstream_sources": [
        {
          "dataset_id": "amazon-berkeley-objects",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#amazon-berkeley-objects",
          "dataset_name": "Amazon Berkeley Objects (ABO)"
        }
      ],
      "description": "Identify a product type"
    },
    {
      "source": "DecisionBench/COM-3",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which industry is this company in?",
      "upstream_sources": [
        {
          "dataset_id": "edgar-corpus-10k-business",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#edgar-corpus-10k-business",
          "dataset_name": "EDGAR-CORPUS (10-K Item 1) with SEC EDGAR SIC codes"
        }
      ],
      "description": "Identify a company industry"
    },
    {
      "source": "DecisionBench/DAT-1",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which SQL answers the question?",
      "upstream_sources": [
        {
          "dataset_id": "spider",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#spider",
          "dataset_name": "Spider (dev set)"
        }
      ],
      "description": "Select the SQL query that answers a question"
    },
    {
      "source": "DecisionBench/DAT-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which cell answers the question?",
      "upstream_sources": [
        {
          "dataset_id": "wikitablequestions",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#wikitablequestions",
          "dataset_name": "WikiTableQuestions (test set)"
        }
      ],
      "description": "Locate the table cell that answers a question"
    },
    {
      "source": "DecisionBench/DAT-4",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What does this column hold?",
      "upstream_sources": [
        {
          "dataset_id": "congress-legislators",
          "selected_questions": 2,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#congress-legislators",
          "dataset_name": "congress-legislators (legislators-current.csv)"
        },
        {
          "dataset_id": "federal-register",
          "selected_questions": 3,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#federal-register",
          "dataset_name": "Federal Register documents API"
        },
        {
          "dataset_id": "fiscal-data",
          "selected_questions": 1,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#fiscal-data",
          "dataset_name": "U.S. Treasury Fiscal Data"
        },
        {
          "dataset_id": "fivethirtyeight",
          "selected_questions": 4,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#fivethirtyeight",
          "dataset_name": "FiveThirtyEight data repository"
        }
      ],
      "description": "Identify the meaning of a data column"
    },
    {
      "source": "DecisionBench/DOC-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Decision, action item, or neither?",
      "upstream_sources": [
        {
          "dataset_id": "ami",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#ami",
          "dataset_name": "AMI Meeting Corpus"
        }
      ],
      "description": "Classify meeting text as a decision, action item, or neither"
    },
    {
      "source": "DecisionBench/DOC-3",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which summary is this meeting's?",
      "upstream_sources": [
        {
          "dataset_id": "ami",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#ami",
          "dataset_name": "AMI Meeting Corpus"
        }
      ],
      "description": "Select the summary corresponding to a meeting"
    },
    {
      "source": "DecisionBench/ENG-1",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which file must change?",
      "upstream_sources": [
        {
          "dataset_id": "swe-bench-verified",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#swe-bench-verified",
          "dataset_name": "SWE-bench Verified"
        }
      ],
      "description": "Locate the file that needs to change to resolve an issue"
    },
    {
      "source": "DecisionBench/ENG-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What kind of change is this?",
      "upstream_sources": [
        {
          "dataset_id": "commitpackft",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#commitpackft",
          "dataset_name": "CommitPackFT"
        }
      ],
      "description": "Classify the type of a code change"
    },
    {
      "source": "DecisionBench/ENG-3",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which weakness class is this?",
      "upstream_sources": [
        {
          "dataset_id": "nvd",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#nvd",
          "dataset_name": "NVD (National Vulnerability Database)"
        }
      ],
      "description": "Identify a vulnerability weakness class"
    },
    {
      "source": "DecisionBench/ENG-4",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Live secret or placeholder?",
      "upstream_sources": [
        {
          "dataset_id": "creddata",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#creddata",
          "dataset_name": "Samsung CredData"
        }
      ],
      "description": "Distinguish a real secret from a placeholder"
    },
    {
      "source": "DecisionBench/ENG-5",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which message describes this diff?",
      "upstream_sources": [
        {
          "dataset_id": "commitpackft",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#commitpackft",
          "dataset_name": "CommitPackFT"
        }
      ],
      "description": "Select a commit message for a code diff"
    },
    {
      "source": "DecisionBench/FIN-1",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which part of the 10-K?",
      "upstream_sources": [
        {
          "dataset_id": "edgar-corpus",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#edgar-corpus",
          "dataset_name": "EDGAR-CORPUS"
        }
      ],
      "description": "Identify a section of a 10-K filing"
    },
    {
      "source": "DecisionBench/FIN-2",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What does this 8-K report?",
      "upstream_sources": [
        {
          "dataset_id": "sec-edgar-8k",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#sec-edgar-8k",
          "dataset_name": "SEC EDGAR 8-K filings"
        }
      ],
      "description": "Identify the event reported in an 8-K filing"
    },
    {
      "source": "DecisionBench/FIN-4",
      "selected_questions": 10,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 10
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What does this number report?",
      "upstream_sources": [
        {
          "dataset_id": "finer-139",
          "selected_questions": 10,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#finer-139",
          "dataset_name": "FiNER-139"
        }
      ],
      "description": "Identify the financial concept represented by a number"
    },
    {
      "source": "DecisionBench/LEG-1",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Does the NDA say this?",
      "upstream_sources": [
        {
          "dataset_id": "contractnli",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#contractnli",
          "dataset_name": "ContractNLI"
        }
      ],
      "description": "Determine whether an NDA supports a statement"
    },
    {
      "source": "DecisionBench/LEG-2",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What kind of clause is this?",
      "upstream_sources": [
        {
          "dataset_id": "cuad",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#cuad",
          "dataset_name": "CUAD (Contract Understanding Atticus Dataset) v1"
        }
      ],
      "description": "Identify a contract clause type"
    },
    {
      "source": "DecisionBench/LEG-3",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Is this term unfair, and how?",
      "upstream_sources": [
        {
          "dataset_id": "lexglue-unfair-tos",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#lexglue-unfair-tos",
          "dataset_name": "UNFAIR-ToS (LexGLUE)"
        }
      ],
      "description": "Identify an unfair terms-of-service clause and its category"
    },
    {
      "source": "DecisionBench/LEG-4",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Does the contract cover this?",
      "upstream_sources": [
        {
          "dataset_id": "cuad",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#cuad",
          "dataset_name": "CUAD (Contract Understanding Atticus Dataset) v1"
        }
      ],
      "description": "Determine whether a contract includes a specified clause"
    },
    {
      "source": "DecisionBench/PRD-1",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which headline won?",
      "upstream_sources": [
        {
          "dataset_id": "upworthy-archive",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#upworthy-archive",
          "dataset_name": "Upworthy Research Archive (exploratory packages)"
        }
      ],
      "description": "Predict which headline achieved the higher click-through rate"
    },
    {
      "source": "DecisionBench/PRD-2",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Major, minor or patch?",
      "upstream_sources": [
        {
          "dataset_id": "changelogs-bsd",
          "selected_questions": 2,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#changelogs-bsd",
          "dataset_name": "BSD-licensed project CHANGELOG.md files"
        },
        {
          "dataset_id": "changelogs-mit",
          "selected_questions": 7,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#changelogs-mit",
          "dataset_name": "MIT-licensed project CHANGELOG.md files"
        }
      ],
      "description": "Classify a version change as major, minor, or patch"
    },
    {
      "source": "DecisionBench/SAF-1",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Attack on the assistant",
      "upstream_sources": [
        {
          "dataset_id": "deepset-prompt-injections",
          "selected_questions": 5,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#deepset-prompt-injections",
          "dataset_name": "deepset/prompt-injections"
        },
        {
          "dataset_id": "in-the-wild-jailbreak-prompts",
          "selected_questions": 4,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#in-the-wild-jailbreak-prompts",
          "dataset_name": "TrustAIRLab/in-the-wild-jailbreak-prompts"
        }
      ],
      "description": "Distinguish prompt injection, jailbreak, and benign requests"
    },
    {
      "source": "DecisionBench/SAF-2",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Toxic comment or not",
      "upstream_sources": [
        {
          "dataset_id": "civil-comments",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#civil-comments",
          "dataset_name": "google/civil_comments"
        }
      ],
      "description": "Identify toxic comments"
    },
    {
      "source": "DecisionBench/SUP-1",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "Which team owns this complaint?",
      "upstream_sources": [
        {
          "dataset_id": "cfpb-complaints",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#cfpb-complaints",
          "dataset_name": "CFPB Consumer Complaint Database"
        }
      ],
      "description": "Route a consumer complaint to the responsible product team"
    },
    {
      "source": "DecisionBench/SUP-2",
      "selected_questions": 9,
      "sector": "domain_decisions",
      "original_splits": {
        "test": 9
      },
      "kinds": {
        "choice": 9
      },
      "immediate_source_url": "https://github.com/atlanai/decision-bench",
      "immediate_source_revision": "6fed2cd4c3608b070e649180796ccaef3d020a23",
      "immediate_corpus": "bench-v4",
      "task_name": "What is the complaint about?",
      "upstream_sources": [
        {
          "dataset_id": "cfpb-complaints",
          "selected_questions": 9,
          "source_catalog_url": "https://github.com/atlanai/decision-bench/blob/6fed2cd4c3608b070e649180796ccaef3d020a23/data/SOURCES.md#cfpb-complaints",
          "dataset_name": "CFPB Consumer Complaint Database"
        }
      ],
      "description": "Identify the issue raised by a consumer complaint"
    },
    {
      "source": "Fashion200k-category",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "test": 15,
        "dev": 15
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/Marqo/fashion200k",
      "description": "Clothing image classification using category1 labels."
    },
    {
      "source": "GQA",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "dev": 27,
        "test": 3
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/lmms-lab-encoder/GQA",
      "description": "Yes/no visual question answering."
    },
    {
      "source": "GoEmotions",
      "selected_questions": 50,
      "sector": "multiselect",
      "original_splits": {
        "dev": 50
      },
      "kinds": {
        "multi_choice": 50
      },
      "upstream_url": "https://huggingface.co/datasets/google-research-datasets/go_emotions",
      "description": "Multi-select emotion classification from the simplified validation split."
    },
    {
      "source": "JevAdvBench",
      "selected_questions": 13,
      "sector": "adversarial",
      "original_splits": {
        "test": 6,
        "dev": 7
      },
      "kinds": {
        "choice": 13
      },
      "upstream_url": "https://huggingface.co/datasets/Hangtao/JevAdvBench",
      "description": "Decision questions with adversarial input perturbations and upstream reference labels."
    },
    {
      "source": "MMMU",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "test": 25,
        "dev": 5
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/MMMU/MMMU",
      "description": "Multidisciplinary image-and-text multiple-choice questions."
    },
    {
      "source": "MMMU-Pro",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "test": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/MMMU/MMMU_Pro",
      "description": "Advanced multidisciplinary image-and-text multiple-choice questions."
    },
    {
      "source": "MathVista",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "test": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/AI4Math/MathVista",
      "description": "Candidate-based mathematics questions with images."
    },
    {
      "source": "MultiRC",
      "selected_questions": 50,
      "sector": "multiselect",
      "original_splits": {
        "dev": 50
      },
      "kinds": {
        "multi_choice": 50
      },
      "upstream_url": "https://huggingface.co/datasets/aps/super_glue",
      "description": "Reading comprehension with multiple correct answer candidates."
    },
    {
      "source": "OpenJev-games/community-diversity-v2/games/pit_stop",
      "selected_questions": 22,
      "sector": "games",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "choice": 11,
        "score": 11
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Racing pit-stop and service-plan selection."
    },
    {
      "source": "OpenJev-games/community-diversity-v2/games/racing_overtake",
      "selected_questions": 22,
      "sector": "games",
      "original_splits": {
        "dev": 22
      },
      "kinds": {
        "choice": 11,
        "score": 11
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Racing overtake candidate selection."
    },
    {
      "source": "OpenJev-games/snake-v1",
      "selected_questions": 10,
      "sector": "games",
      "original_splits": {
        "test": 8,
        "dev": 2
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Snake action decisions."
    },
    {
      "source": "OpenJev-games/tic_tac_toe-v1",
      "selected_questions": 10,
      "sector": "games",
      "original_splits": {
        "test": 8,
        "dev": 2
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Tic-tac-toe move decisions."
    },
    {
      "source": "OpenJev-games/tile_platformer-v1",
      "selected_questions": 10,
      "sector": "games",
      "original_splits": {
        "dev": 2,
        "test": 8
      },
      "kinds": {
        "choice": 10
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Platformer action decisions."
    },
    {
      "source": "OpenJev-games/trex_runner-v1",
      "selected_questions": 6,
      "sector": "games",
      "original_splits": {
        "test": 4,
        "dev": 2
      },
      "kinds": {
        "choice": 6
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Runner action decisions."
    },
    {
      "source": "OpenJev-games/vizdoom-basic-v1",
      "selected_questions": 20,
      "sector": "games",
      "original_splits": {
        "test": 13,
        "dev": 7
      },
      "kinds": {
        "choice": 10,
        "score": 10
      },
      "immediate_source_url": "https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1",
      "upstream_url": "https://github.com/Zefan-Cai/Open-Jev",
      "description": "Action selection from ViZDoom basic states."
    },
    {
      "source": "PhishNChips",
      "selected_questions": 43,
      "sector": "safety",
      "original_splits": {
        "test": 43
      },
      "kinds": {
        "choice": 43
      },
      "upstream_url": "https://huggingface.co/datasets/AreLit/PhishNChips",
      "description": "Phishing classification from email text and a URL."
    },
    {
      "source": "ScienceQA-full",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "dev": 16,
        "test": 14
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/derek-thomas/ScienceQA",
      "description": "Science multiple-choice questions with images.",
      "declared_upstream_license": "cc-by-nc-sa-4.0",
      "license_reference_url": "https://github.com/lupantech/ScienceQA/blob/main/LICENSE-DATA"
    },
    {
      "source": "THIS-THAT-complex",
      "selected_questions": 19,
      "sector": "spatial_rules",
      "original_splits": {
        "dev": 8,
        "test": 11
      },
      "kinds": {
        "choice": 19
      },
      "upstream_url": "https://huggingface.co/datasets/limberc/this-that-complex-decisions",
      "description": "Decisions involving complex rules and conflicting constraints."
    },
    {
      "source": "THIS-THAT-spatial",
      "selected_questions": 6,
      "sector": "spatial_rules",
      "original_splits": {
        "dev": 1,
        "test": 5
      },
      "kinds": {
        "choice": 6
      },
      "upstream_url": "https://huggingface.co/datasets/limberc/this-that-spatial-bench",
      "description": "Spatial relation decisions."
    },
    {
      "source": "TextVQA",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "dev": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/lmms-lab/textvqa",
      "description": "Yes/no image questions involving text."
    },
    {
      "source": "VQAv2",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "dev": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/lmms-lab/VQAv2",
      "description": "Yes/no visual question answering."
    },
    {
      "source": "VirtualHome-executable-plan",
      "selected_questions": 13,
      "sector": "embodied_actions",
      "original_splits": {
        "dev": 13
      },
      "kinds": {
        "choice": 13
      },
      "upstream_url": "https://github.com/xavierpuigf/virtualhome",
      "description": "Select the demonstrated next program action from a task and previous actions."
    },
    {
      "source": "VizWiz",
      "selected_questions": 30,
      "sector": "visual",
      "original_splits": {
        "dev": 30
      },
      "kinds": {
        "choice": 30
      },
      "upstream_url": "https://huggingface.co/datasets/lmms-lab/VizWiz-VQA",
      "description": "Yes/no questions about user-submitted images."
    },
    {
      "source": "arc_challenge",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "choice": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/allenai/ai2_arc",
      "upstream_config": "ARC-Challenge",
      "description": "Science knowledge and reasoning multiple-choice questions",
      "declared_upstream_license": "cc-by-sa-4.0"
    },
    {
      "source": "banking77",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "choice": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/mteb/banking77",
      "upstream_config": null,
      "description": "Routing banking requests to 77 customer intent classes",
      "declared_upstream_license": "cc-by-4.0 (PolyAI; mirror: mit)"
    },
    {
      "source": "boolq",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "noul": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/google/boolq",
      "upstream_config": "default",
      "description": "Yes/no reading comprehension with a supplied passage",
      "declared_upstream_license": "cc-by-sa-3.0"
    },
    {
      "source": "chaosnli",
      "sector": "language_knowledge_judgment",
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/metaeval/chaos-mnli-ambiguity",
      "description": "Natural-language inference with distributions from 100 annotations per example; selected MNLI portion.",
      "declared_upstream_license": "cc-by-nc-4.0",
      "license_reference": "https://raw.githubusercontent.com/easonnie/ChaosNLI/master/LICENSE",
      "selected_questions": 23,
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "choice": 23
      }
    },
    {
      "source": "civil_comments",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "noul": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/google/civil_comments",
      "upstream_config": "default",
      "description": "Comment toxicity and human annotation distributions",
      "declared_upstream_license": "cc0-1.0"
    },
    {
      "source": "clinc150",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "choice": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/clinc/clinc_oos",
      "upstream_config": "plus",
      "description": "150 intent classes plus out-of-scope classification",
      "declared_upstream_license": "cc-by-3.0"
    },
    {
      "source": "fever_evidence",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "noul": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/copenlu/fever_gold_evidence",
      "upstream_config": null,
      "description": "Claim support or refutation with supplied evidence",
      "declared_upstream_license": "cc-by-sa-3.0"
    },
    {
      "source": "helpsteer2_helpfulness",
      "selected_questions": 45,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 45
      },
      "kinds": {
        "score": 45
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/nvidia/HelpSteer2",
      "upstream_config": null,
      "description": "Ordinal assessment of response helpfulness",
      "declared_upstream_license": "cc-by-4.0",
      "license_url": "https://creativecommons.org/licenses/by/4.0/"
    },
    {
      "source": "helpsteer2_verbosity",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "score": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/nvidia/HelpSteer2",
      "upstream_config": null,
      "description": "Ordinal assessment of response verbosity",
      "declared_upstream_license": "cc-by-4.0"
    },
    {
      "source": "massive",
      "selected_questions": 23,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 23
      },
      "kinds": {
        "choice": 23
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/mteb/amazon_massive_intent",
      "upstream_config": "en",
      "description": "Assistant intent classification using the English configuration",
      "declared_upstream_license": "cc-by-4.0 (Amazon; mirror: apache-2.0)"
    },
    {
      "source": "measuring_hate_speech",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "score": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/ucberkeley-dlab/measuring-hate-speech",
      "upstream_config": "default",
      "description": "Hate-speech levels and annotator distributions",
      "declared_upstream_license": "cc-by-4.0"
    },
    {
      "source": "mmlu",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "choice": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/cais/mmlu",
      "upstream_config": "all",
      "description": "Multiple-choice knowledge questions across academic subjects",
      "declared_upstream_license": "mit"
    },
    {
      "source": "mnli",
      "sector": "language_knowledge_judgment",
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/nyu-mll/glue",
      "description": "Multi-genre natural-language inference.",
      "declared_upstream_license": "Source-dependent; the MultiNLI authors explicitly permit modification and redistribution for typical machine-learning uses.",
      "license_reference": "https://cims.nyu.edu/~sbowman/multinli/paper.pdf",
      "selected_questions": 22,
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "choice": 22
      }
    },
    {
      "source": "paws",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "noul": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/google-research-datasets/paws",
      "upstream_config": "labeled_final",
      "description": "Semantic equivalence despite high lexical overlap",
      "declared_upstream_license": "other (Google PAWS, free for research and commercial use)"
    },
    {
      "source": "sms_spam",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "noul": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/ucirvine/sms_spam",
      "upstream_config": "plain_text",
      "description": "SMS spam classification",
      "declared_upstream_license": "cc-by-4.0",
      "license_reference_url": "https://archive.ics.uci.edu/dataset/228/sms%2Bspam%2Bcollection"
    },
    {
      "source": "sst5",
      "sector": "language_knowledge_judgment",
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/SetFit/sst5",
      "description": "Sentence-level five-class sentiment assessment from Stanford Sentiment Treebank.",
      "declared_upstream_license": "No explicit redistribution license identified in the inspected original project page and repository README.",
      "license_reference": "https://nlp.stanford.edu/sentiment/",
      "selected_questions": 22,
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "score": 22
      }
    },
    {
      "source": "strategyqa_closed",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "noul": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/ChilleD/StrategyQA",
      "upstream_config": null,
      "description": "Knowledge-based yes/no questions without supporting facts",
      "declared_upstream_license": "mit"
    },
    {
      "source": "strategyqa_grounded",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "noul": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/ChilleD/StrategyQA",
      "upstream_config": null,
      "description": "Yes/no questions with supporting facts",
      "declared_upstream_license": "mit"
    },
    {
      "source": "stsb",
      "selected_questions": 22,
      "sector": "language_knowledge_judgment",
      "original_splits": {
        "test": 22
      },
      "kinds": {
        "score": 22
      },
      "immediate_source_url": "https://huggingface.co/datasets/Praveenrajus/jev-bench",
      "upstream_url": "https://huggingface.co/datasets/sentence-transformers/stsb",
      "upstream_config": "default",
      "description": "Ordinal sentence similarity assessment",
      "declared_upstream_license": "cc-by-sa-4.0 (STS Benchmark)"
    }
  ]
}