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

    Neuronpedia

    INDEX
    Explanations

    separation

    np_max-act · gemini-2.0-flash

    <start> projects 0.26 may 0.46 come 1.37 off 2.82 during 1.58 bicycle 0.22 <end> <start> slip 1.91 from 2.79 when 1.58 <end> Explanation of neuron 4 behavior: the main thing this neuron does is find language describing parts detaching or slipping off.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    mechanical components or devices related to attachment and coupling in machinery.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    bubble
    -0.07
    cot
    -0.07
    墨
    -0.06
    _bag
    -0.06
    en
    -0.06
    NUMBER
    -0.06
    ubble
    -0.06
    quests
    -0.06
    itchen
    -0.06
     loaf
    -0.06
    POSITIVE LOGITS
     claimed
    0.07
    /**↵↵
    0.07
     카
    0.06
     |↵↵
    0.06
     Nur
    0.06
    ...↵↵
    0.06
    	Array
    0.06
     QIcon
    0.06
    *******/↵↵
    0.06
    CBC
    0.06
    Act Density 0.021%

    No Known Activations