© 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. 24-GEMMASCOPE-2-TRANSCODER-262K
    4. 179321
    Prev
    Next
    INDEX
    Explanations

    * `MAX_ACTIVATING_TOKENS`: `true`, `true`, `True`, `True`, `True`, `TRUE`, `TRUE`, `True`, `true`, `true`* `TOKENS_AFTER_MAX_ACTIVATING_TOKEN`: `}}`, `,`, `},`, `)`, `,`, `,`, `,`The `MAX_ACTIVATING_TOKENS` list is identical (case-insenstitive) and consists solely of "true". The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` list consists of common list/dictionary delimiters (`,`, `}`, `)`).This suggests the neuron is highly sensitive to the boolean value `true` appearing in contexts that are typically followed by closing delimiters or list separators.Looking at `TOP_ACTIVATING_TEXTS`, examples include:1.**boolean true**

    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_24_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
     música
    0.36
     Dune
    0.36
     Chinese
    0.35
     China
    0.33
     degrees
    0.33
    ">
    0.32
    也可以
    0.32
     cát
    0.32
     संज्ञा
    0.31
     eines
    0.31
    POSITIVE LOGITS
    䅡
    0.36
    करी
    0.34
    нали
    0.33
    genome
    0.33
     ड्राइ
    0.32
     村
    0.32
    ätten
    0.32
    गन
    0.31
     совместно
    0.31
    timepoint
    0.31
    Activations Density 0.001%

    No Known Activations