© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129096
    Prev
    Next
    INDEX
    Explanations

    The neuron is simply tracking where you are in the text – it fires more strongly on tokens that occur farther into the document (i.e. it encodes absolute token position).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    quoted speech

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     and
    -1.77
    hline
    -1.41
     tačiau
    -1.27
     также
    -1.23
     useAppContext
    -1.18
    จึง
    -1.13
     نیز
    -1.12
     schreef
    -1.12
     versioni
    -1.09
     BSEB
    -1.05
    POSITIVE LOGITS
     I
    1.95
     it
    1.73
     we
    1.66
     there
    1.61
     wasn
    1.45
     you
    1.43
     everything
    1.40
     they
    1.38
     It
    1.38
    に行って
    1.28
    Activations Density 0.003%

    No Known Activations