DarwinAnim8or commited on
Commit
cf748d5
·
verified ·
1 Parent(s): 9f4a858

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +157 -3
README.md CHANGED
@@ -1,3 +1,157 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ task_categories:
4
+ - text-classification
5
+ tags:
6
+ - text-moderation
7
+ language:
8
+ - en
9
+ - de
10
+ - fr
11
+ - es
12
+ - it
13
+ - sv
14
+ - fi
15
+ - pl
16
+ - cs
17
+ - lv
18
+ - zh
19
+ - ja
20
+ - ko
21
+ - ru
22
+ - uk
23
+ - be
24
+ - kk
25
+ ---
26
+
27
+ # Text-Moderation-Multilingual
28
+
29
+ A comprehensive multilingual text moderation dataset combining multiple high-quality sources for training robust content moderation classifiers.
30
+
31
+ ## Dataset Summary
32
+
33
+ This dataset aggregates text moderation data from multiple sources to create a large-scale, diverse training corpus for content moderation systems. It includes text samples labeled across multiple harmful content categories, supporting both multilingual and English-specific moderation use cases.
34
+
35
+ **Total Size:** ~1.7M entries
36
+ **Languages:** Multilingual (primary focus on English)
37
+ **Task:** Multi-label text classification for content moderation
38
+
39
+ ## Dataset Structure
40
+
41
+ ### Data Fields
42
+
43
+ - `prompt` (string): The input text to be classified
44
+ - `S` (int): Sexual content (0 = safe, 1 = harmful)
45
+ - `H` (int): Hate speech (0 = safe, 1 = harmful)
46
+ - `V` (int): Violence (0 = safe, 1 = harmful)
47
+ - `HR` (int): Harassment (0 = safe, 1 = harmful)
48
+ - `SH` (int): Self-harm (0 = safe, 1 = harmful)
49
+ - `S3` (int): Sexual content involving minors (0 = safe, 1 = harmful)
50
+ - `H2` (int): Hate speech (alternative labeling) (0 = safe, 1 = harmful)
51
+ - `V2` (int): Violence (alternative labeling) (0 = safe, 1 = harmful)
52
+
53
+ ### Data Splits
54
+
55
+ - **Train:** 1459350 samples
56
+ - **Validation:** 162150 samples
57
+
58
+ *Note: Split created with 90/10 train/validation ratio using random seed 42*
59
+
60
+ ## Source Datasets
61
+
62
+ This dataset combines and harmonizes data from:
63
+
64
+ - **[ifmain's multilingual dataset](https://huggingface.co/datasets/ifmain/text-moderation-02-multilingual)** - Multilingual moderation examples
65
+ - **[OpenAI's English evaluation dataset](https://huggingface.co/datasets/mmathys/openai-moderation-api-evaluation)** - High-quality English evaluation samples
66
+ - **[ifmain's English dataset](https://huggingface.co/datasets/ifmain/text-moderation-01)** - English moderation examples
67
+
68
+ ## Intended Use
69
+
70
+ ### Primary Use Cases
71
+ - Training text moderation classifiers
72
+ - Benchmarking content moderation systems
73
+ - Research into automated content moderation
74
+ - Multi-label classification model development
75
+
76
+ ### Out-of-Scope Uses
77
+ - This dataset is **not intended** for any purpose other than training content moderation systems
78
+ - Should not be used to generate harmful content
79
+ - Not suitable for general text classification tasks outside of moderation
80
+
81
+ ## Considerations for Using the Data
82
+
83
+ ### Content Warning
84
+ This dataset contains examples of harmful content including hate speech, harassment, violence, and other potentially disturbing material. Users should exercise appropriate caution when working with this data.
85
+
86
+ ### Bias and Limitations
87
+ - The dataset reflects the biases present in the source datasets
88
+ - Content moderation standards may vary across different platforms and cultures
89
+ - Label consistency across merged datasets may vary
90
+ - Primarily English-focused despite multilingual components
91
+
92
+ ### Ethical Considerations
93
+ - This dataset should only be used to improve content moderation and safety systems
94
+ - Researchers and developers should implement appropriate safeguards when working with this data
95
+ - The goal is to reduce harmful content online, not to amplify it
96
+
97
+ ## Example Usage
98
+
99
+ ```python
100
+ from datasets import load_dataset
101
+
102
+ # Load the dataset
103
+ dataset = load_dataset("KoalaAI/Text-Moderation-Multilingual")
104
+
105
+ # Access splits
106
+ train_data = dataset["train"]
107
+ val_data = dataset["validation"]
108
+
109
+ # Example entry
110
+ print(train_data[0])
111
+ # {
112
+ # 'prompt': 'Example text...',
113
+ # 'S': 0, 'H': 0, 'V': 0, 'HR': 0,
114
+ # 'SH': 0, 'S3': 0, 'H2': 0, 'V2': 0
115
+ # }
116
+ ```
117
+
118
+ ## Dataset Creation
119
+
120
+ ### Curation Process
121
+ 1. Source datasets were identified and downloaded
122
+ 2. Data was harmonized to use consistent labeling schema
123
+ 3. Entries were merged and deduplicated where appropriate
124
+ 4. Train/validation split was created using stratified sampling
125
+
126
+ ### Quality Control
127
+ - Labels were preserved from original high-quality sources
128
+ - Data integrity checks were performed during merging process
129
+ - Consistent schema applied across all entries
130
+
131
+ ## License
132
+
133
+ Please refer to the licenses of the individual source datasets:
134
+ - Check ifmain datasets for their respective licensing terms
135
+ - OpenAI evaluation dataset licensing applies to that portion
136
+ - Usage should comply with all source dataset requirements
137
+
138
+ ## Citation
139
+
140
+ If you use this dataset, please cite the original source datasets:
141
+
142
+ ```bibtex
143
+ @misc{text-moderation-large,
144
+ title={Text-Moderation-Multilingual: A Multilingual Text Moderation Dataset},
145
+ author={[KoalaAI]},
146
+ year={2025},
147
+ note={Aggregated from ifmain's and OpenAI's moderation datasets}
148
+ }
149
+ ```
150
+
151
+ ## Contact
152
+
153
+ For questions about this dataset compilation, please open an issue on this repository.
154
+
155
+ ---
156
+
157
+ **Disclaimer:** This dataset is provided for research and safety purposes only. Users are responsible for ensuring ethical use and compliance with applicable laws and regulations.