license: mit

fant-koala-modedist

description

fant-koala-modedist is a pseudo-labeled moderation dataset created from akaruineko/fantastic-offensive using the KoalaAI/Text-Moderation model.

The dataset stores the teacher model's predicted moderation labels together with their probabilities, making it suitable for experiments with multi-label classification, pseudo-labeling, and knowledge distillation.

pipeline

akaruineko/fantastic-offensive
            ↓
   KoalaAI/Text-Moderation
            ↓
     label probabilities
            ↓
     fant-koala-modedist
            ↓
      FTAN multi-label

fields

• text — original text from the source dataset
• moderation_results — predicted moderation labels and their probabilities

Example:

{
  "text": "example text",
  "moderation_results": [
    {"label": "LABEL_A", "probability": 0.91},
    {"label": "LABEL_B", "probability": 0.07},
    {"label": "LABEL_C", "probability": 0.02}
  ]
}

intended use

This dataset is primarily intended for:

• training multi-label moderation models
• teacher-student experiments
• knowledge distillation
• pseudo-labeling experiments
• research on automated content moderation

source

Original dataset: akaruineko/fantastic-offensive

Teacher model: KoalaAI/Text-Moderation

note

The labels are model-generated pseudo-labels, not manually verified annotations. The teacher model may therefore produce incorrect or biased predictions.

This dataset should be treated as an automatically generated training resource rather than a ground-truth benchmark.