ariannap22/collectiveaction_sft_annotated_only_v6_prompt_v6_p100_synthetic_balanced_more_layered
The ariannap22/collectiveaction_sft_annotated_only_v6_prompt_v6_p100_synthetic_balanced_more_layered model is an 8 billion parameter fine-tuned LLama3 variant developed by ariannap22. This model is specifically designed to detect levels of participation in collective action from text, serving as the second step in a layered classification approach. It excels at classifying social media comments into categories like 'Problem-Solution', 'Call-to-Action', 'Intention', and 'Execution' based on predefined criteria related to collective action problems.
Loading preview...
Model Overview
This model, developed by ariannap22, is a fine-tuned LLama3 variant with 8 billion parameters, specifically engineered for detecting levels of participation in collective action from textual data. It represents the second step in a two-stage classification process, following an initial binary detection of participation expression using a separate model.
Key Capabilities
- Collective Action Participation Detection: Classifies text into specific levels of participation in collective action, such as 'Problem-Solution', 'Call-to-Action', 'Intention', and 'Execution'.
- Layered Approach: Designed to be used after a preliminary model has identified the binary presence of participation expression.
- Contextual Understanding: Utilizes detailed definitions and criteria for each participation level, enabling nuanced classification of social media comments.
Use Cases
- Social Media Analysis: Ideal for researchers and analysts studying collective action, social movements, and activism by categorizing how individuals express their involvement or intent.
- Sentiment and Engagement Monitoring: Can be integrated into systems that monitor public discourse to understand the nature and intensity of collective action-related discussions.
Framework
The model's underlying framework and methodology are detailed in the paper "Extracting Participation in Collective Action from Social Media" by Pera and Aiello (2025), which provides useful code snippets and theoretical context.