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

    Neuronpedia

    INDEX
    Explanations

    clauses connecting contrast reasons

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     tegens
    -1.08
    Cuar
    -1.06
    文分享
    -0.98
    Probl
    -0.98
    —¿
    -0.97
     měl
    -0.96
    ądze
    -0.96
     этого
    -0.96
     fracaso
    -0.95
    Muz
    -0.94
    POSITIVE LOGITS
     even
    1.31
     only
    1.19
     especially
    1.18
     though
    0.99
     but
    0.96
    只
    0.96
    even
    0.94
     either
    0.93
     because
    0.91
     both
    0.91
    Act Density 0.098%

    No Known Activations