© 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. 24-GEMMASCOPE-2-RES-16K
    7. 2900
    Prev
    Next
    INDEX
    Explanations

    This neuron fires on format‐string and code‐style tokens—particularly percent‐placeholders (like %n, %t, %A) and their surrounding punctuation.

    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_24_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
    <0xE2>
    0.59
     expatri
    0.56
     parties
    0.55
     ‎
    0.54
     KYC
    0.54
     personnel
    0.54
     également
    0.54
     persone
    0.54
     investors
    0.53
     họ
    0.53
    POSITIVE LOGITS
    把
    0.64
     بچہ
    0.59
    就把
    0.59
     调用
    0.58
    thorace
    0.57
     डॉक्यूमेंट्री
    0.57
     ٹھ
    0.57
     rectMode
    0.56
    indx
    0.56
     把
    0.56
    Activations Density 0.002%

    No Known Activations