© 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. 32-GEMMASCOPE-2-TRANSCODER-262K
    4. 235342
    Prev
    Next
    INDEX
    Explanations

    signs for businesses(Rationale: MAX_ACTIVATING_TOKENS includes "(", "**", and words like "Open", "Eat". TOKENS_AFTER_MAX_ACTIVATING_TOKEN includes text like "Please", "Everything", "All", "Dis", which are common in descriptive text or signage. TOP_ACTIVATING_TEXTS heavily features descriptions of establishments, signs, menus, and offerings. TOP_POSITIVE_LOGITS has words like 'east', 'eastbound', 'intersection', 'gas' which can be common in directions or place names. Combining these, a strong pattern emerges around signs indicating places of business or service.)

    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_32_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
    Интернет
    0.44
    ऑनलाइन
    0.41
    Text
    0.40
     Интернет
    0.39
    Online
    0.39
    Internet
    0.38
    Adapt
    0.38
     온라인
    0.38
    Compra
    0.38
    互联网
    0.38
    POSITIVE LOGITS
     east
    0.39
     sister
    0.38
     continuer
    0.38
     intersection
    0.38
    appreci
    0.38
     gas
    0.37
     granny
    0.37
    intersection
    0.37
    cartes
    0.37
     eastbound
    0.37
    Activations Density 0.000%

    No Known Activations