File size: 3,878 Bytes
d5cca0c b30358f d5cca0c abc899d d5cca0c ef69456 f632b52 b7ce4cd f632b52 b7ce4cd ef69456 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 |
---
dataset_info:
features:
- name: uid
dtype: string
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: label
dtype:
class_label:
names:
'0': entailment
'1': neutral
'2': contradiction
- name: reason
dtype: string
splits:
- name: train
num_bytes: 44720719
num_examples: 100459
- name: validation
num_bytes: 663148
num_examples: 1200
- name: test
num_bytes: 657586
num_examples: 1200
download_size: 15202058
dataset_size: 46041453
task_categories:
- text-classification
language:
- en
tags:
- NLI
size_categories:
- 100K<n<1M
---
# The Adversarial Natural Language Inference (ANLI)
- Source: https://huggingface.co/datasets/anli
- Num examples:
- 100,459 (train)
- 1,200 (validation)
- 1,200 (test)
- Language: English
```python
from datasets import load_dataset
load_dataset("vietgpt/anli_r3_en")
```
- Format for NLI task
```python
def preprocess(sample):
premise = sample['premise']
hypothesis = sample['hypothesis']
label = sample['label']
if label == 0:
label = "entailment"
elif label == 1:
label = "neutral"
else:
label = "contradiction"
return {'text': f'<|startoftext|><|premise|> {premise} <|hypothesis|> {hypothesis} <|label|> {label} <|endoftext|>'}
"""
<|startoftext|><|premise|> TOKYO, Dec 18 (Reuters) - Japan’s Shionogi & Co said on Tuesday that it has applied to health regulators in the United States, Canada and Europe for approval of its HIV drug Dolutegravir. Shionogi developed Dolutegravir with a Viiv Healthcare, an AIDS drug joint venture between GlaxoSmithKline and Pfizer, in exchange for its rights to the drug. <|hypothesis|> The article was written on December 18th. <|label|> entailment <|endoftext|>
"""
```
- Format for Rationale task
```python
def preprocess_rationale(sample):
premise = sample['premise']
hypothesis = sample['hypothesis']
rationale = sample['reason']
return {'text': f'<|startoftext|><|premise|> {premise} <|hypothesis|> {hypothesis} <|rationale|> {rationale} <|endoftext|>'}
"""
<|startoftext|><|premise|> TOKYO, Dec 18 (Reuters) - Japan’s Shionogi & Co said on Tuesday that it has applied to health regulators in the United States, Canada and Europe for approval of its HIV drug Dolutegravir. Shionogi developed Dolutegravir with a Viiv Healthcare, an AIDS drug joint venture between GlaxoSmithKline and Pfizer, in exchange for its rights to the drug. <|hypothesis|> The article was written on December 18th. <|rationale|> TOKYO, Dec 18 (Reuters) is when the article was written as it states in the first words of the sentence <|endoftext|>
"""
```
- Format for GPT-3
```python
def preprocess_gpt3(sample):
premise = sample['premise']
hypothesis = sample['hypothesis']
label = sample['label']
if label == 0:
output = f'\n<|correct|> True\n<|incorrect|> False\n<|incorrect|> Neither'
elif label == 1:
output = f'\n<|correct|> Neither\n<|incorrect|> True\n<|incorrect|> False'
else:
output = f'\n<|correct|> False\n<|incorrect|> True\n<|incorrect|> Neither'
return {'text': f'<|startoftext|> anli 2: {premise} <|question|> {hypothesis}\nTrue, False, or Neither? <|answer|> {output} <|endoftext|>'}
"""
<|startoftext|> anli 2: TOKYO, Dec 18 (Reuters) - Japan’s Shionogi & Co said on Tuesday that it has applied to health regulators in the United States, Canada and Europe for approval of its HIV drug Dolutegravir. Shionogi developed Dolutegravir with a Viiv Healthcare, an AIDS drug joint venture between GlaxoSmithKline and Pfizer, in exchange for its rights to the drug. <|question|> The article was written on December 18th.
True, False, or Neither? <|answer|>
<|correct|> True
<|incorrect|> False
<|incorrect|> Neither <|endoftext|>
"""
``` |