© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11875
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly responds to “How to” instructional phrases, i.e. leading question or tutorial prompts starting with “How to.”

    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
     zwar
    0.71
     ró
    0.65
    ming
    0.62
     VS
    0.62
    द
    0.60
     understand
    0.59
    </b>
    0.59
     irgendwie
    0.59
     nix
    0.57
    </strong>
    0.56
    POSITIVE LOGITS
    aries
    0.61
     Strengthen
    0.57
     टीच
    0.57
     ಸದಸ್ಯ
    0.56
    ？
    0.55
     TextAppearance
    0.55
     تقویت
    0.55
     strengthened
    0.55
     Pray
    0.54
    됩
    0.54
    Activations Density 0.132%

    No Known Activations