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

    Neuronpedia

    INDEX
    Explanations

    academic texts

    np_max-act · gemini-2.0-flash

    This neuron detects structured outline headings—particularly the numbered modules, lessons, and section identifiers in a training/course index.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     warming
    -0.07
    Agent
    -0.06
    -prop
    -0.06
    IFA
    -0.06
    -H
    -0.06
    �
    -0.06
    BP
    -0.06
    Mini
    -0.06
    character
    -0.06
    utoff
    -0.06
    POSITIVE LOGITS
    ???↵↵
    0.07
    391
    0.07
     그는
    0.07
    );//
    0.07
     appending
    0.07
     byl
    0.06
     Guidelines
    0.06
    .....↵↵
    0.06
     blames
    0.06
     RTE
    0.06
    Act Density 0.135%

    No Known Activations