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

    Neuronpedia

    INDEX
    Explanations

    causal conjunctions

    np_max-act · gemini-2.0-flash

    This neuron detects discourse connectors and logical-transition phrases (e.g. “In other words,” “Now,” “Since,” “Therefore,” etc.) indicating shifts or links in the proof’s argument.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    zug
    -0.07
    :i
    -0.06
    ERAL
    -0.06
    汗
    -0.06
    (hash
    -0.06
    sher
    -0.06
    rical
    -0.06
    喝
    -0.06
     tribe
    -0.06
     Tf
    -0.06
    POSITIVE LOGITS
     Thank
    0.07
    	boolean
    0.07
     sponsoring
    0.06
    	device
    0.06
     Nederland
    0.06
     kaç
    0.06
    _softmax
    0.06
    _SAMPLE
    0.06
     فکی
    0.06
    .fhir
    0.06
    Act Density 0.020%

    No Known Activations