#MODELCitizens
February 16, 2025 at 4:13 AM
The call for participation for the 40th Chaos Communication Congress with the motto „Model Citizen“ has opened.

Find all information here: https://content.events.ccc.de/cfp/40c3/index.en.html

#40c3 #modelcitizen
40C3 Call For Participation
content.events.ccc.de
September 11, 2026 at 6:17 PM
Work in progress pic of my Sylvaneth army.

This model is huge, getting close to complete.

Aim to have her ready for Lost Legion GT this coming weekend.

#1st post
#ageofsigmar
#warhammer
#modelcitizens
#tabletopmodels
November 25, 2024 at 2:01 AM
(1/3) Please check out our new paper with @skgabrie.bsky.social and her amazing students, to appear in #EMNLP2025!

(🚨 Offensive Content Warning)

arxiv.org/abs/2507.05455
ModelCitizens: Representing Community Voices in Online Safety
Automatic toxic language detection is critical for creating safe, inclusive online spaces. However, it is a highly subjective task, with perceptions of toxic language shaped by community norms and liv...
arxiv.org
August 26, 2025 at 6:12 PM
April 14, 2026 at 3:35 PM
February 21, 2026 at 6:42 PM
(2/3) Toxicity detection is shaped by norms, identity, & context, which existing approaches overlook. Enter MODELCITIZENS: a new dataset designed to address this.
✔️ 6.8K posts, 40K annotations across diverse groups
✔️ Context-augmented scenarios
✔️ New fine-tuned models that beat GPT-4o-mini by 5.5%
August 26, 2025 at 6:14 PM
Models struggle to capture this nuance:
🤖 GPT-o4-mini, OpenAI Moderation, Gemini-2.0-Flash reach ~63% accuracy on ModelCitizens
🚀 Fine-tuned LLAMACITIZEN-8B and GEMMACITIZEN-12B outperform them by +5.5% on in-distribution tests, better aligning with community perspectives. (5/6🧵)
April 22, 2026 at 4:36 PM
💡 ModelCitizens introduces a new approach:
📝 6,822 social media posts
📊 40K toxicity annotations
🌍 8 historically targeted identity groups
💬 Context-augmented posts generated with LLMs
It keeps ingroup and outgroup perspectives separate. (3/6🧵)
April 22, 2026 at 4:35 PM