© 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. 40-GEMMASCOPE-2-RES-16K
    7. 11277
    Prev
    Next
    INDEX
    Explanations

    almost always

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_40_width_16k_l0_medium
    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
     frequ
    0.46
    Frequent
    0.44
     Frequent
    0.43
    잦
    0.40
    حديث
    0.39
     frequent
    0.38
     অনেকে
    0.37
     vielf
    0.36
    *}$
    0.34
    +}$
    0.34
    POSITIVE LOGITS
     always
    3.58
     siempre
    3.30
     всегда
    3.28
     sempre
    3.20
    always
    3.20
     Always
    3.09
    Always
    3.08
     завжди
    3.08
     alltid
    2.98
     ALWAYS
    2.98
    Activations Density 0.655%

    No Known Activations