stereo

Riferimenti:

interfrase

Utilizzare il comando seguente per caricare questo set di dati in TFDS:

ds = tfds.load('huggingface:stereoset/intersentence')
  • Descrizione :
Stereoset is a dataset that measures stereotype bias in language models. Stereoset consists of 17,000 sentences that
measures model preferences across gender, race, religion, and profession.
  • Licenza : CC BY-SA 4.0
  • Versione : 1.0.0
  • Divide :
Diviso Esempi
'validation' 2123
  • Caratteristiche :
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "bias_type": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentences": {
        "feature": {
            "sentence": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "labels": {
                "feature": {
                    "label": {
                        "num_classes": 4,
                        "names": [
                            "anti-stereotype",
                            "stereotype",
                            "unrelated",
                            "related"
                        ],
                        "names_file": null,
                        "id": null,
                        "_type": "ClassLabel"
                    },
                    "human_id": {
                        "dtype": "string",
                        "id": null,
                        "_type": "Value"
                    }
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "gold_label": {
                "num_classes": 3,
                "names": [
                    "anti-stereotype",
                    "stereotype",
                    "unrelated"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

intrafrase

Utilizzare il comando seguente per caricare questo set di dati in TFDS:

ds = tfds.load('huggingface:stereoset/intrasentence')
  • Descrizione :
Stereoset is a dataset that measures stereotype bias in language models. Stereoset consists of 17,000 sentences that
measures model preferences across gender, race, religion, and profession.
  • Licenza : CC BY-SA 4.0
  • Versione : 1.0.0
  • Divide :
Diviso Esempi
'validation' 2106
  • Caratteristiche :
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "bias_type": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentences": {
        "feature": {
            "sentence": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "labels": {
                "feature": {
                    "label": {
                        "num_classes": 4,
                        "names": [
                            "anti-stereotype",
                            "stereotype",
                            "unrelated",
                            "related"
                        ],
                        "names_file": null,
                        "id": null,
                        "_type": "ClassLabel"
                    },
                    "human_id": {
                        "dtype": "string",
                        "id": null,
                        "_type": "Value"
                    }
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "gold_label": {
                "num_classes": 3,
                "names": [
                    "anti-stereotype",
                    "stereotype",
                    "unrelated"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}