© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 526
    Prev
    Next
    INDEX
    Explanations

    in summary, in essence, in short, in conclusion

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on sentence-initial discourse markers—especially prepositions and similar transition words like “In,” “By,” “With,” “For,” “As,” and “At.”

    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
    /
    0.43
    <unused312>
    0.34
    เจ้า
    0.34
    。
    0.34
    <unused105>
    0.33
                          
    0.33
     /.
    0.33
     ибо
    0.33
     /
    0.32
     دیا۔
    0.32
    POSITIVE LOGITS
    此同时
    0.57
    clusion
    0.50
    oltre
    0.50
     hindsight
    0.48
    credibly
    0.48
    此之外
    0.47
     November
    0.46
    quiries
    0.46
    spiration
    0.46
    volvement
    0.46
    Activations Density 1.378%

    No Known Activations