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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 9338
    Prev
    Next
    INDEX
    Explanations

    indicates dévelo fournisseur

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron spikes on common English suffixes and clitic endings (e.g. –ing, –ly, –ity, ’s) rather than on root words themselves.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     was
    -1.46
    为了
    -1.39
    不同的
    -1.38
    通过
    -1.34
     which
    -1.32
     This
    -1.31
    不同
    -1.30
     For
    -1.30
     How
    -1.24
    一些
    -1.18
    POSITIVE LOGITS
     dévelo
    1.39
    婠
    1.31
     asust
    1.31
    鲧
    1.22
    一辈子
    1.20
    都没
    1.19
     Virol
    1.18
    indicates
    1.18
     fournisseur
    1.17
    菒
    1.17
    Activations Density 0.181%

    No Known Activations