© 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. 2469
    Prev
    Next
    INDEX
    Explanations

    The neuron is activated by occurrences of the word “bubble” (in any form or compound).

    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
     biał
    0.60
     embeddings
    0.57
     เด้อ
    0.55
     destinados
    0.55
    $$\
    0.53
     velvety
    0.53
     Vamos
    0.52
     congru
    0.52
     قاف
    0.52
     zewnętrzne
    0.50
    POSITIVE LOGITS
    asList
    0.57
    typeof
    0.57
    இந்
    0.55
    🧍
    0.54
    ((
    0.52
    pip
    0.51
     учу
    0.51
    ้ย
    0.50
    GOOGLE
    0.50
    Tail
    0.50
    Activations Density 0.014%

    No Known Activations