FrenchCastle/sexology-v4

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Nov 29, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

FrenchCastle/sexology-v4 is a 7 billion parameter instruction-tuned causal language model developed by a French public-health body, specialized in providing reliable, caring, and evidence-based sexual health information in French. Fine-tuned from Mistral-7B-Instruct-v0.3, it excels at answering general sexual-health questions and directing users to professional care, with a 4096-token context length. Its primary differentiator is its focus on factual reliability, a non-judgemental tone, and safe behavior within the sensitive domain of sexual health.

Loading preview...

Sexo-FR: A Specialized French Sexual Health Information Model

Sexo-FR is a 7 billion parameter instruction-tuned causal language model, developed as part of a French public-health initiative. It is specifically designed to provide reliable, caring, and evidence-based sexual health information in plain French, while consistently guiding users to qualified professionals and dedicated helplines. The model is an informational and educational tool, not a medical device, and does not provide diagnoses or replace healthcare consultations.

Key Capabilities

  • Factual Reliability: Content is aligned with recognized public-health and clinical guidance.
  • Caring & Inclusive Tone: Emphasizes a non-judgemental and inclusive approach.
  • Safe Behavior: Refuses sexually explicit content, recognizes distress, and signposts professional resources.
  • Comprehensive Topics: Covers contraception, STIs, consent, sexual and reproductive health, relationships, sexual well-being, and inclusive information on gender identity and sexual orientation.

Good For

  • Powering French-language assistants for general sexual-health questions.
  • Deployment in chatbots or Q&A features on official public-health websites.
  • Serving as an FAQ-assistance layer or an internal drafting aid for human experts.
  • Further adaptation for specific public-health campaigns or integration into retrieval-augmented generation (RAG) systems for authoritative, updated knowledge.