© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 27-GEMMASCOPE-2-TRANSCODER-262K
    4. 197613
    Prev
    Next
    INDEX
    Explanations

    The `MAX_ACTIVATING_TOKENS` list is strange as it's full of `(`. This is not a typical sequence of tokens. It might indicate a problem with the input data or a neuron that reacts to unusual structural patterns.The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` list yields words commonly found in programming, especially concerning data types and API endpoints: `string`, `int`, `FromBody`, `I`, `data`, `item`.The `TOP_POSITIVE_LOGITS` are diverse. Some might relate to programming (`Firebase`, `HANDLER`), others are foreign words or seemingly unrelated.The `TOP_ACTIVATING_TEXTS` show code snippets.`FromBody` or `IFormFile` in C#

    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
    ettin
    0.43
    جمع
    0.40
     discret
    0.38
    yanka
    0.38
     intermission
    0.38
     interment
    0.37
     वर्गों
    0.37
     dirig
    0.37
     spel
    0.36
    owy
    0.36
    POSITIVE LOGITS
    maktan
    0.39
     gratuito
    0.38
     гран
    0.36
    Firebase
    0.34
     মাহমুদ
    0.34
     tumorigen
    0.34
     warrior
    0.33
    ER
    0.33
     dijadikan
    0.33
    HANDLER
    0.33
    Activations Density 0.000%

    No Known Activations