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

    Neuronpedia

    INDEX
    Explanations

    low-rank trainable matrices

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    GRESS
    0.44
    داخل
    0.38
    Aub
    0.37
    PERTY
    0.37
    zież
    0.36
     urldecode
    0.36
    句话
    0.36
     mercy
    0.36
     Perc
    0.36
     అందు
    0.36
    POSITIVE LOGITS
     matrices
    0.49
    Rank
    0.49
    adapter
    0.48
     adapter
    0.47
     adapters
    0.47
     layers
    0.46
     Rank
    0.45
     Adapter
    0.44
    matrices
    0.43
     matrix
    0.42
    Act Density 0.017%

    No Known Activations