© 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
    2. Gemma-3-27B-IT
    3. 40-GEMMASCOPE-2-TRANSCODER-262K
    4. 28304
    Prev
    Next
    INDEX
    Explanations

    This neuron appears to be identifying locations, geographical features, and descriptive words associated with them. It seems to focus on regions and specific natural elements.Here are some of the patterns observed:1. **Geographical Locations:** The presence of "Oregon", "Idaho", "western North America", "western US", "Alaska", and names like "Lassen Volcanic National Park" strongly suggest the neuron is sensitive to place names.2. **Natural Features/Environments:** Words like "desert", "lava", "forests", "lakes", "meadow", "pine", "river" point towards an interest in natural landscapes and vegetation.3. **Descriptive Adjectives/Verbs:** Terms like "showcasing", "lush", "wet", "drier", "rugged" indicate it might be capturing descriptive elements related to these locations.4. **Contextual Phrases:** Seeing phrases like "rain shadow desert" or descriptions of climates further reinforces the idea of geographical and environmental context.Given these observations, a concise explanation could be:**western us geography and environments**

    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_40_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
    Attachments
    0.38
    attachments
    0.36
     Attachment
    0.35
     attachments
    0.35
    xcsche
    0.35
    Так
    0.35
    triangleq
    0.35
    gabe
    0.34
     किसको
    0.34
     άνθρω
    0.34
    POSITIVE LOGITS
     meadow
    0.47
     Klam
    0.46
     Desch
    0.46
     bend
    0.44
     Bend
    0.43
     Redmond
    0.43
     lava
    0.41
     Ondo
    0.41
     Gerardo
    0.40
     Mod
    0.40
    Activations Density 0.001%

    No Known Activations