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

    Neuronpedia

    INDEX
    Explanations

    can

    np_max-act · gemini-2.0-flash

    expressions or phrases related to language comprehension and communication.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron strongly activates on occurrences of the modal verb “can” (and variants expressing ability), i.e. statements of what the assistant is able to do.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     advantages
    -0.07
     accommodations
    -0.07
     analyze
    -0.07
    observer
    -0.07
     vyu
    -0.07
     belongings
    -0.06
    Nach
    -0.06
    Bo
    -0.06
    perator
    -0.06
    izace
    -0.06
    POSITIVE LOGITS
    *****↵↵
    0.06
    ][-
    0.06
    浩
    0.06
     Colum
    0.06
     могу
    0.06
    _MAXIMUM
    0.06
     Güney
    0.06
     gồm
    0.06
    으면
    0.06
     yoktu
    0.06
    Act Density 0.039%

    No Known Activations