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

    Neuronpedia

    INDEX
    Explanations

    significantly improve, better understand, effectively biasing

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     هذا
    1.59
     этом
    1.50
     this
    1.46
     tomto
    1.45
     questo
    1.43
     tohoto
    1.33
     dieser
    1.27
     này
    1.27
     этому
    1.27
     это
    1.24
    POSITIVE LOGITS
    㳑
    1.15
    妽
    1.09
    🈯
    1.05
    所需的
    1.01
    📂
    0.98
    💪
    0.97
    ic
    0.96
    <unused549>
    0.94
    羍
    0.93
    荿
    0.93
    Act Density 0.552%

    No Known Activations