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

    Neuronpedia

    INDEX
    Explanations

    Gemma, CF, file, RDS, interview

    np_acts-logits-general · gemini-2.5-flash-lite

    the neuron detects prominent technical topic keywords — especially acronyms, product/model names, and domain-specific terms.

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31
    New Auto-Interp
    Negative Logits
    t
    0.90
    c
    0.84
    in
    0.82
    k
    0.80
    l
    0.73
    time
    0.72
    ol
    0.71
    al
    0.71
    b
    0.71
    se
    0.70
    POSITIVE LOGITS
    :
    0.71
     políticas
    0.61
     désir
    0.56
     zákon
    0.56
     humoral
    0.56
     deverá
    0.55
     onSuccess
    0.54
     isOn
    0.54
     hood
    0.53
     político
    0.53
    Act Density 0.130%

    No Known Activations