© 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. 19898
    Prev
    Next
    INDEX
    Explanations

    describing states or qualities

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

    The main thing this neuron does is detect very common, high-frequency tokens (e.g. common stop words and other frequent short words).

    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
     ---
    -1.66
     anvä
    -1.53
     -----
    -1.48
    the
    -1.45
     each
    -1.44
     -
    -1.44
    liyor
    -1.41
     thier
    -1.41
    起码
    -1.38
    5
    -1.38
    POSITIVE LOGITS
    </i>
    1.61
     菇
    1.59
     Câ
    1.55
     Anglais
    1.46
     ché
    1.41
    <bos>
    1.38
     chiffres
    1.38
     he
    1.36
    ienie
    1.36
     bateaux
    1.34
    Activations Density 0.013%

    No Known Activations