Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    capturing long-range dependencies

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     encoder
    0.45
    isotropic
    0.40
     Jit
    0.39
     Wür
    0.39
     Encode
    0.39
     KMeans
    0.38
    encoder
    0.38
     информационной
    0.38
    decoder
    0.38
    ResponseWriter
    0.38
    POSITIVE LOGITS
     dependencies
    0.65
     relationships
    0.64
     Dependencies
    0.59
    relationships
    0.57
     hierarchy
    0.54
     relationship
    0.54
     Relationships
    0.53
    dependencies
    0.53
    Dependencies
    0.52
    Hier
    0.52
    Act Density 0.049%

    No Known Activations