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

    Neuronpedia

    INDEX
    Explanations

    conversation/quotes

    np_max-act · gemini-2.0-flash

    markers that denote the start of the assistant’s reply in chat-formatted conversations.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     san
    -0.07
     avant
    -0.07
    _name
    -0.07
     cach
    -0.06
     sob
    -0.06
    requ
    -0.06
     tenía
    -0.06
     MÜ
    -0.06
    lifetime
    -0.06
     distortion
    -0.06
    POSITIVE LOGITS
     %%↵
    0.07
     ناحیه
    0.07
     Therm
    0.07
    に関する
    0.06
    hen
    0.06
    .“
    0.06
    HEN
    0.06
    	getline
    0.06
                               ↵
    0.06
    itemap
    0.06
    Act Density 0.092%

    No Known Activations