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

    Neuronpedia

    INDEX
    Explanations

    hallucinations and fabrications

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     cautioned
    0.73
     criticised
    0.72
    唏
    0.70
     avoided
    0.69
     evas
    0.69
     lame
    0.68
     controversial
    0.67
     criticized
    0.66
     sluggish
    0.64
     dismal
    0.63
    POSITIVE LOGITS
     believing
    2.46
     belief
    2.20
     believe
    2.15
    believe
    1.94
     believes
    1.89
    Believe
    1.88
    belief
    1.84
     Believe
    1.83
    以為
    1.78
     Belief
    1.75
    Act Density 0.195%

    No Known Activations