Your Algorithm Comes to Facebook Reels

Meta is bringing Instagram’s Your Algorithm feature to Facebook, giving users a way to influence the Reels recommendations that appear in their feeds. The feature is designed to make recommendations more responsive to people’s interests, rather than leaving them with little say over what the platform suggests.

On Instagram, Your Algorithm lets users view and manage topics that shape the Reels they see. Meta says the feature will now be available on Facebook as well, extending that approach to another of its social platforms. Users can use it to signal which subjects they want to see more or less often, helping guide the recommendations shown to them.

The change is part of Meta’s effort to give people more control over their Reels experience. Instead of relying only on the platform’s automated recommendations, Facebook users will have a way to express their preferences directly. The feature focuses specifically on Reels recommendations, offering a tool to adjust the topics reflected in that stream. Meta has not framed it as a replacement for recommendation systems; rather, it gives users an additional way to shape the content those systems surface.

How Users Can Shape Recommendations

Facebook’s Your Algorithm controls let people tell the platform which interests they want to see more or less of in Reels. The aim is to give users more influence over the videos recommended to them, rather than leaving recommendations entirely to Facebook’s system.

Users can review topics Facebook thinks they are interested in and adjust those preferences. They can ask to see more content related to a topic or less of it, helping shape what appears in their Reels feed. These controls are designed to make recommendations more relevant to each person’s stated interests.

Your Algorithm also offers a way to understand why certain videos are being shown. By viewing and changing interest preferences, users can signal what they want their recommendations to include or avoid. The controls do not mean people choose every individual video they see, but they provide a direct way to guide the topics Facebook uses when making recommendations.

Meta’s Increased Use of LLMs

Meta is using large language models (LLMs) more often to decide which content people see on Facebook. The company says these models are helping it better understand and rank content, including posts and videos in Reels. That builds on Meta’s recommendation systems, which use signals about what people watch, like, share, and comment on to suggest content they may find interesting.

The increased use of LLMs is part of a broader effort to make recommendations more responsive to people’s interests. Meta says the models can help its systems interpret content and user activity more effectively, rather than relying only on simpler signals. Recommendations can also be shaped by users: people can provide feedback on what they want to see more or less of, and that feedback can influence what appears in their feeds.

Meta’s approach therefore combines automated ranking with user choices. LLMs help assess and rank content, while the recommendation system draws on people’s activity and feedback to tailor what it shows. The company says it will continue using these tools to improve recommendations across Facebook.

What the Changes Mean for Facebook

Facebook’s Reels recommendations are becoming more responsive to both what people do and what they say they want to see. Watching, liking, sharing, and commenting on videos can help shape recommendations, while the “Interested” and “Not interested” options give users a more direct way to influence what appears. These signals work alongside Meta’s systems for selecting and ranking Reels, rather than leaving recommendations entirely to manual choices.

Meta’s increased use of large language models adds another layer to that process. The company says LLMs can help its systems understand content and user interests more deeply, and improve how recommendations are tailored. In combination, behavioral signals, explicit feedback, and AI tools may help Facebook surface videos that better match each person’s interests. But recommendations are still shaped by the platform’s ranking systems: user input can guide them, while Meta’s models help interpret that input and decide which Reels to show.