© 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. 34-GEMMASCOPE-RES-131K
    4. 103030
    Prev
    Next
    INDEX
    Explanations

    `{` `}`

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

    The neuron consistently picks out isolated numeric tokens used as formal identifiers (e.g. reference numbers, version numbers, years, or section IDs).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    чника
    -0.88
    Reference
    -0.76
     reference
    -0.74
    REFERENCE
    -0.68
    Eyes
    -0.68
    reference
    -0.66
     terci
    -0.66
    anhänger
    -0.66
     sement
    -0.65
     Holi
    -0.64
    POSITIVE LOGITS
    ocyto
    0.81
    締切
    0.71
    doch
    0.71
    бей
    0.65
    Todo
    0.65
     rundt
    0.65
    UnknownFields
    0.64
    吠
    0.64
    哪些
    0.63
     caliper
    0.63
    Activations Density 0.064%

    No Known Activations