metadata/croissant.json

{
  "@context": {
    "@language": "en",
    "@vocab": "https://schema.org/",
    "citeAs": "cr:citeAs",
    "column": "cr:column",
    "conformsTo": "dct:conformsTo",
    "cr": "http://mlcommons.org/croissant/",
    "rai": "http://mlcommons.org/croissant/RAI/",
    "data": {
      "@id": "cr:data",
      "@type": "@json"
    },
    "dataType": {
      "@id": "cr:dataType",
      "@type": "@vocab"
    },
    "dct": "http://purl.org/dc/terms/",
    "examples": {
      "@id": "cr:examples",
      "@type": "@json"
    },
    "extract": "cr:extract",
    "field": "cr:field",
    "fileProperty": "cr:fileProperty",
    "fileObject": "cr:fileObject",
    "fileSet": "cr:fileSet",
    "format": "cr:format",
    "includes": "cr:includes",
    "isLiveDataset": "cr:isLiveDataset",
    "jsonPath": "cr:jsonPath",
    "key": "cr:key",
    "md5": "cr:md5",
    "parentField": "cr:parentField",
    "path": "cr:path",
    "recordSet": "cr:recordSet",
    "references": "cr:references",
    "regex": "cr:regex",
    "repeated": "cr:repeated",
    "replace": "cr:replace",
    "sc": "https://schema.org/",
    "separator": "cr:separator",
    "source": "cr:source",
    "subField": "cr:subField",
    "transform": "cr:transform"
  },
  "@type": "sc:Dataset",
  "name": "seqBench",
  "description": "\nSeqBench is a tunable benchmark designed to probe and analyze sequential reasoning \ncapabilities in language models. This compact version contains instances with varied \ncomplexity. Each instance provides:\n- 'context': NLP problem.\n- 'completion': Solution.\n- 'complexity_parameters': A dictionary with L, B, N.\n- 'instance_metadata': Maze dimensions and agent/target names.\n- 'structural_details': Rich structural information (as a JSON string) including room mappings, \n  adjacency lists, door/key details, and canonical facts for the underlying base maze.\n",
  "conformsTo": "http://mlcommons.org/croissant/1.0",
  "citeAs": "@misc{anonymous2025seqbench,\n  author    = {Anonymous Submission},\n  title     = {SeqBench: A Tunable Benchmark to Quantify Sequential Reasoning Limits of LLMs},\n  year      = {2025},\n  publisher = {Proceedings of the Conference on Empirical Methods in Natural Language Processing},\n  note      = {Special Theme: Interdisciplinary Recontextualization of NLP},\n  comment   = {Dataset accessible at https://huggingface.co/datasets/emnlp-submission/seqBench}\n}",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "url": "https://huggingface.co/datasets/emnlp-submission/seqBench",
  "distribution": [
    {
      "@type": "cr:FileObject",
      "@id": "seqbench-jsonl-gz-file",
      "name": "seqbench-jsonl-gz-file",
      "description": "The main benchmark data file in gzipped JSONL format.",
      "contentUrl": "https://huggingface.co/datasets/emnlp-submission/seqBench/resolve/main/seqBench_compact.jsonl.gz",
      "encodingFormat": "application/gzip",
      "sha256": "2b58a2f7b65def3445f2572726d5ff30d5ecab9cc841fffcb57a8bceb937b0c3"
    }
  ],
  "recordSet": [
    {
      "@type": "cr:RecordSet",
      "@id": "seqbench-records",
      "name": "seqbench-records",
      "description": "Individual instances from the SeqBench benchmark (extracted from gzipped JSONL).",
      "field": [
        {
          "@type": "cr:Field",
          "@id": "context",
          "name": "context",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "column": "context"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "completion",
          "name": "completion",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "column": "completion"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "logical_depth_L",
          "name": "logical_depth_L",
          "dataType": "sc:Integer",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "jsonPath": "complexity_parameters.logical_depth_L"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "backtracking_count_B",
          "name": "backtracking_count_B",
          "dataType": "sc:Integer",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "jsonPath": "complexity_parameters.backtracking_count_B"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "noise_ratio_N",
          "name": "noise_ratio_N",
          "dataType": "sc:Float",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "jsonPath": "complexity_parameters.noise_ratio_N"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "instance_id",
          "name": "instance_id",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "column": "instance_id"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "full_instance_metadata",
          "name": "full_instance_metadata",
          "description": "Full instance metadata as a JSON string (contains maze_rows, maze_cols, agent_name, target_name).",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "column": "instance_metadata"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "structural_details_json",
          "name": "structural_details_json",
          "description": "Full structural details as a JSON string.",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "seqbench-jsonl-gz-file"
            },
            "extract": {
              "column": "structural_details"
            }
          }
        }
      ]
    }
  ]
}