© 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. 12239
    Prev
    Next
    INDEX
    Explanations

    This neuron activates on out‐of‐vocabulary tokens—chiefly numeric tokens and foreign (non-Latin) script tokens.

    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
    تهي
    0.55
    Optim
    0.55
    Semantic
    0.55
    http
    0.54
    Recognition
    0.54
    о
    0.53
    くと
    0.53
    Fisher
    0.53
    Discussion
    0.52
    শায়
    0.52
    POSITIVE LOGITS
     geldi
    0.68
     TikTok
    0.64
     geliyor
    0.63
     🥰
    0.63
    .​
    0.63
     Comes
    0.62
     Tiktok
    0.61
     oluşan
    0.60
     kommer
    0.60
     trekking
    0.60
    Activations Density 0.008%

    No Known Activations