© 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-3-27B-IT
    5. 7-GEMMASCOPE-2-TRANSCODER-262K
    7. 35061
    Prev
    Next
    INDEX
    Explanations

    dialogue opening

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_7_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     plupart
    0.39
    ра
    0.39
    is
    0.37
    s
    0.37
    t
    0.36
    রা
    0.35
    the
    0.34
    स
    0.34
    ام
    0.33
    f
    0.33
    POSITIVE LOGITS
     to
    0.42
     if
    0.39
     andere
    0.36
     إذا
    0.35
     such
    0.35
     don
    0.35
    endente
    0.33
     do
    0.33
     নিজের
    0.33
     או
    0.33
    Activations Density 0.016%

    No Known Activations