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

    Neuronpedia

    INDEX
    Explanations

    dates and years

    np_max-act-logits · gemini-2.5-flash-lite

    This neuron detects numeric tokens—especially years, dates, and other numbers.

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31

    years and dates

    np_max-act · gpt-5-miniTriggered by @yooniel31
    New Auto-Interp
    Negative Logits
     next
    3.84
    next
    3.52
    Next
    3.13
     Next
    2.98
     próxima
    2.91
     NEXT
    2.81
     nächste
    2.81
     nächsten
    2.72
    NEXT
    2.72
     próximo
    2.70
    POSITIVE LOGITS
    今年も
    0.52
    今年は
    0.44
     今年
    0.41
    今年的
    0.39
    今年の
    0.38
    今年
    0.38
    ِينَ
    0.37
    銖
    0.35
    昨年
    0.35
    今天要
    0.35
    Act Density 0.010%

    No Known Activations