Coding
PromptBeginner5 minmarkdown
Nano Banana Pro
Agent skill for nano-banana-pro
7
Coding
FrenchToxicityPrompts
**Name**: FrenchToxicityPrompts
promptBeginner5 min to valuemarkdown
0 views
Jan 15, 2026
Sign in to like and favorite skills
**Name**: FrenchToxicityPrompts
Sign in to like and favorite skills
Name: FrenchToxicityPrompts
Overview: A dataset of 50,000 naturally occurring French prompts and their continuations, annotated with toxicity scores from a widely used toxicity classifier, designed to evaluate toxicity in generated content by language models.
Data Type: text
Domains:
Languages:
Resources:
Goal: To evaluate and mitigate toxicity in French text generated by language models.
Target Audience:
Tasks:
Limitations: The dataset covers exclusively French data, and the toxicity scores are dependent on Perspective API.
Source: Lélu, a French written dialogue dataset extracted from Reddit’s public dataset.
Size: 50,000 prompts and continuations
Format: JSON
Annotation: Annotated using the Perspective API for various toxicity attributes.
Methods:
Metrics:
Calculation: Metrics are calculated based on the scores provided by the Perspective API for toxicity attributes.
Interpretation: Higher toxicity metrics indicate greater potential for the model to generate toxic content.
Validation: Dataset was manually examined for a selection of samples to ensure the validity of toxicity annotations.
Risk Categories:
Atlas Risks: No specific atlas risks defined
Demographic Analysis: N/A
Potential Harm: N/A
Privacy And Anonymity: The dataset contains explicit content and harmful language which requires careful handling.
Data Licensing: Not Applicable
Consent Procedures: Not Applicable
Compliance With Regulations: Not Applicable