What problem it solves

A personalized feed must select relevant posts, rank them, and determine whether they can be shown. The implementation combines posts from followed and unfollowed accounts with filtering, transformer-based ranking, labeling, and visibility controls.

Key features

• Personalized candidate retrieval — The feed gathers candidates from accounts the viewer follows and from accounts the viewer does not follow. Retrieval mechanisms include Phoenix, SimClusters, and a shared list of popular recent posts.

  Source: README, phoenix

• Transformer-based ranking — Phoenix uses the viewer's recent engagement history to predict actions on candidate posts. Predicted values are combined with configurable weights and other scoring adjustments to order the feed.

  Source: README, home-mixer

• Separate visibility filtering — Visibility decisions are made after ranking using viewer actions, account and post labels, and other contextual inputs. Posts may be allowed, given a notice, placed behind an interstitial, or dropped.

  Source: README, visibility-filtering

• Content and account labeling — Models examine posts, media, and accounts before applying rules that produce labels. Those labels are stored and later used by visibility filtering to determine how content can appear.

  Source: README, scarecrow

• Configurable pipeline and transparency — Pipeline stages and their defaults are controlled through configuration, while Under the Hood exposes aggregate information about labels that can limit account or post visibility. The transparency tool also reports certain legally required visibility limitations.

  Source: README, under-the-hood

Who it may suit

This may suit developers and researchers who want to study feed candidate generation, ranking, labeling, and visibility-filtering systems. It may also suit community members interested in how For You feed decisions and transparency reporting are implemented.

Topics to discuss

• How should transformer-based action predictions be evaluated for individual feed relevance?

• Which visibility decisions should be surfaced in user-facing transparency reports?

• How can out-of-network retrieval balance discovery with safety and filtering constraints?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/xai-org/x-algorithm#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.