© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 27-GEMMASCOPE-2-TRANSCODER-262K
    4. 118419
    Prev
    Next
    INDEX
    Explanations

    Based on the provided lists, the neuron appears to be identifying specific formatting markers and sizing indicators commonly found in web design code or structured text.* **MAX_ACTIVATING_TOKENS**: `sm`, `md`, `lg`, `For`, `/`, `Note`, `*`, `px`* **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: `: `, `Books`, `May`, `)**`Observations:* `sm`, `md`, `lg`, `px` are common CSS class prefixes for responsive design breakpoints or pixel values. They are frequently followed by `:` (e.g., `md:gap-6`, `md:text-lg`, `text-sm`).* `For` is often followed by `Books` in the context of providing advice or categories (`* For Books:`).* `Note` is followed by `:` indicating an explanation or annotation.* `*` is followed by a delimiter `:` and then content (`wall`).The common theme is finding specific tokens that act as delimiters or initiators for structured information, particularly related to sizing/styling and categorization.Thus, a good explanation would be:**sizing prefixes and textual markers**sizing prefixes and textual markers

    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_27_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
    Rt
    0.40
    𝒮
    0.40
     EVs
    0.35
     تقسيم
    0.35
    ᱥ
    0.35
     окна
    0.35
    सि
    0.34
     গর
    0.34
     Anas
    0.34
    すれば
    0.34
    POSITIVE LOGITS
    corp
    0.42
    ventures
    0.41
    ome
    0.41
    etano
    0.41
    non
    0.40
     कैसी
    0.40
    charisma
    0.39
    annie
    0.38
    berto
    0.38
     وفر
    0.38
    Activations Density 0.000%

    No Known Activations