RSTZZZZ commited on
Commit
9a1b5d2
Β·
verified Β·
1 Parent(s): dde2632

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +106 -0
README.md CHANGED
@@ -19,3 +19,109 @@ size_categories:
19
  ---
20
 
21
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
19
  ---
20
 
21
 
22
+ # MLSNT: Multi-Lingual Social Network Toxicity Dataset
23
+
24
+ **MLSNT** is a multi-lingual dataset for toxicity detection created through a large language model-assisted label transfer pipeline. It enables efficient and scalable moderation across languages and platforms, and is built to support span-level and category-specific classification for toxic content.
25
+
26
+ This dataset is introduced in the following paper:
27
+
28
+ > **Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for Resource-Efficient Toxicity Detection**
29
+ > πŸ† Accepted at **KDD 2025**, Applied Data Science Track
30
+
31
+ ---
32
+
33
+ ## 🧩 Overview
34
+
35
+ MLSNT harmonizes 15 publicly available toxicity datasets across **7 languages** using GPT-4o-mini to create consistent binary and fine-grained labels. It is suitable for both training and evaluating toxicity classifiers in multi-lingual, real-world moderation systems.
36
+
37
+ ---
38
+
39
+ ## 🌍 Supported Languages
40
+
41
+ - πŸ‡«πŸ‡· French (`fr`)
42
+ - πŸ‡©πŸ‡ͺ German (`de`)
43
+ - πŸ‡΅πŸ‡Ή Portuguese (`pt`)
44
+ - πŸ‡·πŸ‡Ί Russian (`ru`)
45
+ - πŸ‡¨πŸ‡³ Simplified Chinese (`zh-cn`)
46
+ - πŸ‡ΉπŸ‡Ό Traditional Chinese (`zh-tw`)
47
+ - πŸ‡―πŸ‡΅ Japanese (`ja`)
48
+
49
+ ---
50
+
51
+ ## πŸ—οΈ Construction Method
52
+
53
+ 1. **Source Datasets**
54
+ 15 human-annotated datasets were gathered from `hatespeechdata.com` and peer-reviewed publications.
55
+
56
+ 2. **LLM-Assisted Label Transfer**
57
+ GPT-4o-mini was prompted to re-annotate each instance into a unified label schema. Only examples where human and LLM annotations agreed were retained.
58
+
59
+ 3. **Toxicity Categories**
60
+ Labels are fine-grained categories (e.g., `threat`, `hate`, `harassment`).
61
+
62
+ ---
63
+
64
+ ## πŸ“Š Dataset Statistics
65
+
66
+ | Language | Total Samples | % Discarded | Toxic % (Processed) |
67
+ |----------------------|----------------|-------------|----------------------|
68
+ | German (HASOC, etc.) | ~13,800 | 28–69% | 32–56% |
69
+ | French (MLMA) | ~3,200 | 20% | 94% |
70
+ | Russian | ~14,300 | ~40% | 33–54% |
71
+ | Portuguese | ~21,000 | 20–44% | 26–50% |
72
+ | Japanese | ~2,000 | 10–25% | 17–45% |
73
+ | Chinese (Simplified) | ~34,000 | 29–46% | 48–61% |
74
+ | Chinese (Traditional)| ~65,000 | 37% | ~9% |
75
+
76
+ ---
77
+
78
+ ## πŸ’Ύ Format
79
+
80
+ Each row in the dataset includes:
81
+
82
+ - `full_text`: The original utterance or message.
83
+ - `start_string_index`: A list of start string indices (start positions of toxic spans).
84
+ - `end_string_index`: A list of end string indices (end positions of toxic spans).
85
+ - `category_id`: A list of toxic category IDs (integer values).
86
+ - `final_label`: A list of toxic category names (string values).
87
+ - `min_category_id`: The minimum toxic category ID in the row (used as the primary label).
88
+ - `match_id`: A unique identifier composed of the original dataset name and a row-level ID.
89
+
90
+ ---
91
+
92
+ ## πŸ—‚οΈ Category ID Mapping
93
+
94
+ | ID | Friendly Name |
95
+ |-----|---------------------------------------------|
96
+ | 0 | Non Toxic |
97
+ | 1 | Threats (Life Threatening) |
98
+ | 2 | Minor Endangerment |
99
+ | 3 | Threats (Non-Life Threatening) |
100
+ | 4 | Hate |
101
+ | 5 | Sexual Content / Harassment |
102
+ | 6 | Extremism |
103
+ | 7 | Insults |
104
+ | 8 | Controversial / Potentially Toxic Topic |
105
+
106
+ ---
107
+
108
+ ## πŸ”¬ Applications
109
+
110
+ - Fine-tuning multi-lingual moderation systems
111
+ - Cross-lingual toxicity benchmarking
112
+ - Training span-level and category-specific toxicity detectors
113
+ - Studying LLM label transfer reliability and agreement filtering
114
+
115
+ ---
116
+
117
+ ## πŸ“œ Citation
118
+
119
+ If you use MLSNT in academic work, please cite:
120
+
121
+ ```bibtex
122
+ @inproceedings{yang2025mlsnt,
123
+ title={Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for Resource-Efficient Toxicity Detection},
124
+ author={Zachary Yang and Domenico Tullo and Reihaneh Rabbany},
125
+ booktitle={Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)},
126
+ year={2025}
127
+ }