© 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. 13288
    Prev
    Next
    INDEX
    Explanations

    This neuron specializes in spotting formal, document‐style or legal/official discourse markers and section‐opening phrases (e.g. “NOW, THEREFORE,” “WHEREAS,” “border search,” etc.) that signal the start of clauses or structured legal/administrative text.

    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
    ())){
    0.73
     있는데
    0.69
    ()){
    0.65
     했는데
    0.65
    했는데
    0.64
    还有一个
    0.63
    >(){
    0.62
     इतना
    0.61
    ਾਰ
    0.60
     하는데
    0.60
    POSITIVE LOGITS
     therefore
    5.16
    therefore
    4.72
    Therefore
    4.44
     Therefore
    4.28
     поэтому
    4.28
     derfor
    4.25
    因此
    4.19
     pertanto
    4.06
     därför
    4.03
     hence
    3.88
    Activations Density 0.695%

    No Known Activations