© 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. 90600
    Prev
    Next
    INDEX
    Explanations

    special characters and punctuation

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

    The neuron detects decorative separator lines made up of repeated tilde characters (i.e. long runs of “~” used as horizontal dividers).

    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
     There
    -1.22
    无疑
    -1.16
     different
    -1.16
    のお話
    -1.11
     various
    -1.09
     ziemlich
    -1.09
    ****
    -1.09
    Мария
    -1.09
     This
    -1.09
    Due
    -1.09
    POSITIVE LOGITS
     which
    2.63
     and
    2.22
     but
    2.02
     namely
    1.46
    which
    1.42
    whose
    1.27
     then
    1.24
     lakini
    1.23
     whereas
    1.22
     however
    1.19
    Activations Density 0.101%

    No Known Activations