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

    Neuronpedia

    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    requests for suggestions or recommendations.

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

    This neuron responds to negation phrases that express someone lacking or not having something (e.g. “don’t have one yet”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    -icons
    -0.07
    _train
    -0.07
    	old
    -0.07
    Collider
    -0.06
    Problem
    -0.06
     snap
    -0.06
    thed
    -0.06
     fit
    -0.06
     Olymp
    -0.06
    (Http
    -0.06
    POSITIVE LOGITS
    enefit
    0.07
    iPad
    0.06
     strugg
    0.06
    WORD
    0.06
     Agents
    0.06
    ención
    0.06
    يلا
    0.06
    ريف
    0.06
    edi
    0.06
     울
    0.06
    Act Density 0.215%

    No Known Activations