© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 69144
    Prev
    Next
    INDEX
    Explanations

    capital letters after punctuation

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

    The neuron strongly activates on numbers (digits and numeric tokens) and related punctuation (colons, question marks) as they appear in headings or metadata fields.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     then
    -1.03
     with
    -1.01
    Then
    -1.00
    oid
    -0.98
     those
    -0.94
     there
    -0.92
     for
    -0.91
    两者
    -0.91
     even
    -0.90
    自身的
    -0.89
    POSITIVE LOGITS
    已经
    1.09
    描述
    1.03
     araña
    0.99
     verhind
    0.98
    ЕВ
    0.96
    junge
    0.93
     decidi
    0.93
    kontakt
    0.92
    ukasz
    0.92
    şan
    0.92
    Activations Density 0.228%

    No Known Activations