One key limitation of keyword-based classifiers is that they don’t understand meaning in context. For example, the same word can mean very different things depending on how it’s used—something a simple keyword system might miss. To address this, we can use more advanced tools called language models. These models are trained on vast amounts of text to learn how words, phrases, and ideas typically relate to each other. Put simply for our use case: language models are much better at classifying complicated things such as broad topics (”politics” or “gender”) or tone (“intimidation” or “negative sentiment” or “toxicity”).
While general-purpose language models are powerful, they still need to be adapted to the specific needs of peacebuilding. This process, known as fine-tuning, involves teaching a model to recognize the kinds of content that matter in your work—such as harmful speech, key topics like gender or migration, or early signs of escalation. To support this, we’ve curated a set of models that perform well on these tasks, including some we’ve fine-tuned ourselves and others selected based on their strong performance. These models are ready to use and can help you classify social media posts more accurately and meaningfully—without requiring technical expertise.
When you click “create” in the Classify tab on the Phoenix platform, you will see options to create an author classifier and a keyword text classifier (these two options are described in the previous sections). You will also see options to apply complex models, open source classifiers, to classify text (in posts or comments), with the name of the model and a short description. The models listed below are currently available on Phoenix. Click on the dropdown button for information on how they work and how you can apply them to your data.
<aside> <img src="/icons/search_blue.svg" alt="/icons/search_blue.svg" width="40px" />
Can’t see a model that works for you? We aim to add more classifier models as we grow. Do you have access to a model that you think would help Phoenix? Email [email protected] to let us know and we’ll try to integrate it!
It is possible for peacebuilders to train their own models, here is how we think about this process. We work with a community of developers interested in building models relevant to conflict analysis; you can join our Discord here to find collaborators.
</aside>