© 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. 3-GEMMASCOPE-2-TRANSCODER-262K
    7. 122363
    Prev
    Next
    INDEX
    Explanations

    1. **Identify recurring themes in `MAX_ACTIVATING_TOKENS`**: "Firearms", "Buddhism", "Plateau". These are quite distinct.2. **Look for patterns in `TOKENS_AFTER_MAX_ACTIVATING_TOKEN`**: * "Firearms" is followed by "and". This strongly suggests phrases like "Firearms and explosives" or items listed alongside firearms. * "Buddhism" is followed by verbs like "has", "utilizes", "was". This indicates descriptions *of* Buddhism. * "Plateau" is followed by punctuation like "(" and ",". This suggests a classification or listing context.3.firearms and, buddhism, plateau

    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_3_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
     <
    0.76
    ering
    0.67
     a
    0.66
     e
    0.66
     on
    0.66
     [
    0.65
     _
    0.65
     -
    0.64
    ीकरण
    0.64
     $
    0.63
    POSITIVE LOGITS
    s
    0.72
    0
    0.71
    6
    0.71
    ের
    0.70
    트
    0.66
    ু
    0.64
    timepoint
    0.63
    REPOSITORY
    0.62
    ELEASE
    0.61
    boor
    0.61
    Activations Density 0.000%

    No Known Activations