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

    Neuronpedia

    INDEX
    Explanations

    replicas

    np_max-act · gemini-2.0-flash

    The neuron fires on numeric literals (especially counts like “replicas: 3” or port numbers) in Kubernetes YAML manifests.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     going
    -0.07
     Handlers
    -0.06
     MU
    -0.06
     Cooper
    -0.06
     Morning
    -0.06
     unintended
    -0.06
     Heading
    -0.06
    686
    -0.06
    Browser
    -0.06
    Day
    -0.06
    POSITIVE LOGITS
     replicas
    0.09
     replica
    0.08
    à
    0.07
    RIA
    0.07
    ुम
    0.07
    脸
    0.07
    A
    0.07
    lf
    0.07
    ultipartFile
    0.07
     copy
    0.07
    Act Density 0.001%

    No Known Activations