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

    Neuronpedia

    INDEX
    Explanations

    open data and plans

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    Low
    0.40
    Was
    0.39
    either
    0.39
    both
    0.38
     either
    0.38
    7
    0.37
    all
    0.37
     both
    0.37
    5
    0.36
    8
    0.36
    POSITIVE LOGITS
     downright
    0.50
     prejudices
    0.47
     whatnot
    0.45
     addirittura
    0.44
     importantly
    0.43
     refusal
    0.42
     accusations
    0.42
     frankly
    0.41
     måde
    0.41
     fears
    0.41
    Act Density 0.083%

    No Known Activations