Latent association trees are structured graphs mapping unprompted brand perceptions and secondary concept relationships uncovered by following alternative token branches beneath the surface generation.

graph TD
    Root[Brand or Entity] --> T1[Primary Token 85%]
    Root -.-> T2[Alt Token A 12%]
    Root -.-> T3[Alt Token B 3%]
    T1 --> P1[Top Association]
    T2 --> S1[Latent Sub-Topic 1]
    T2 --> S2[Latent Sub-Topic 2]
    T3 --> S3[Niche Association]