© 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. 12551
    Prev
    Next
    INDEX
    Explanations

    eyes kept drifting

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

    The neuron detects HTML or forum markup tags (angle-bracketed elements like `<div>`, `<BLOCKQUOTE>`, class attributes, etc.).

    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
    വണ
    0.55
    兼
    0.55
     furnish
    0.52
    odial
    0.51
    ំនួន
    0.51
    釆
    0.50
    に変
    0.49
    を決
    0.49
     rescue
    0.49
     expend
    0.48
    POSITIVE LOGITS
     Cassidy
    0.62
    КО
    0.61
     reu
    0.57
     reissue
    0.57
     "&#
    0.55
     Surely
    0.55
     Courthouse
    0.55
     Elijah
    0.54
     Beware
    0.54
    Surely
    0.54
    Activations Density 0.000%

    No Known Activations