© 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. 10-GEMMASCOPE-RES-131K
    7. 1520
    Prev
    Next
    INDEX
    Explanations

    poly(

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

    This neuron activates for special characters and punctuation, particularly parentheses, colons, and various symbolic notations.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron is tuned to spot common function words and punctuation (i.e. high-frequency prepositions, conjunctions, short grammatical words, and symbols).

    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
    ‰
    -1.85
    Ⅴ
    -1.82
    Ⅳ
    -1.82
    tuce
    -1.76
    ichteten
    -1.73
    नलोड
    -1.66
    ígenes
    -1.63
    eacute
    -1.62
     recurrir
    -1.59
    llac
    -1.59
    POSITIVE LOGITS
     also
    2.45
     actually
    1.64
     not
    1.64
    quela
    1.59
     likewise
    1.53
     цього
    1.52
     verdadera
    1.45
     then
    1.44
    </td>
    1.43
    0
    1.42
    Activations Density 0.534%

    No Known Activations