© 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-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 73
    Prev
    Next
    INDEX
    Explanations

    book titles or technical terms

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

    The neuron fires most strongly on multiword, title‐style sequences of capitalized words (e.g. book or article titles).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     с
    0.33
     finalizar
    0.32
     पुअनि
    0.30
     kỳ
    0.29
     yang
    0.29
     allez
    0.29
     ม
    0.28
     actuation
    0.28
     tâm
    0.27
    মঙ্গলবার
    0.27
    POSITIVE LOGITS
     Cookbook
    0.38
    ⨔
    0.37
     Difer
    0.36
     Repar
    0.34
    ￱
    0.34
     Prac
    0.34
     Detalles
    0.34
     Sandals
    0.33
    𝚣
    0.33
     Respect
    0.33
    Activations Density 0.352%

    No Known Activations