© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 3114
    Prev
    Next
    INDEX
    Explanations

    The neuron is primarily activated by numeric tokens or numeric sequences (years, section numbers, dollar amounts, decimals, etc.).

    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
     ಅವನ
    0.61
     (((
    0.54
     Kxe
    0.54
    ]::
    0.51
    ,"$
    0.51
     น้อง
    0.50
     Skin
    0.49
     Schwarz
    0.49
     Tx
    0.49
     TX
    0.49
    POSITIVE LOGITS
    (
    0.79
    (**
    0.63
    **(
    0.55
     मेक
    0.53
     составляет
    0.53
     બનાવવા
    0.52
    égard
    0.50
    עד
    0.50
     интере
    0.50
     имају
    0.49
    Activations Density 0.033%

    No Known Activations