References:
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:liar')
- Description:
LIAR is a dataset for fake news detection with 12.8K human labeled short statements from politifact.com's API, and each statement is evaluated by a politifact.com editor for its truthfulness. The distribution of labels in the LIAR dataset is relatively well-balanced: except for 1,050 pants-fire cases, the instances for all other labels range from 2,063 to 2,638. In each case, the labeler provides a lengthy analysis report to ground each judgment.
- License: Unknown
- Version: 1.0.0
- Splits:
| Split | Examples |
|---|---|
'test' |
1283 |
'train' |
10269 |
'validation' |
1284 |
- Features:
{
"id": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"label": {
"num_classes": 6,
"names": [
"false",
"half-true",
"mostly-true",
"true",
"barely-true",
"pants-fire"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"statement": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"subject": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"speaker": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"job_title": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"state_info": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"party_affiliation": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"barely_true_counts": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"false_counts": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"half_true_counts": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"mostly_true_counts": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"pants_on_fire_counts": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"context": {
"dtype": "string",
"id": null,
"_type": "Value"
}
}