© 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. 128268
    Prev
    Next
    INDEX
    Explanations

    The neuron is essentially detecting document position, firing strongly on tokens that occur around the 400–800-token mark rather than on any linguistic content.

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

    .If

    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
     italiens
    -1.42
     alho
    -1.42
    mètres
    -1.37
    ⚔
    -1.34
     INDIA
    -1.32
     cannes
    -1.31
     FRANCE
    -1.31
     herbe
    -1.31
    ☺☺
    -1.30
    乀
    -1.30
    POSITIVE LOGITS
     that
    1.63
     these
    1.57
     this
    1.54
     other
    1.38
     them
    1.27
    就可以了
    1.26
    完成です
    1.24
     those
    1.20
    或者是
    1.20
     such
    1.20
    Activations Density 0.011%

    No Known Activations