© 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
    3. Gemma-3-27B-IT
    5. 31-GEMMASCOPE-2-TRANSCODER-262K
    7. 163048
    Prev
    Next
    INDEX
    Explanations

    file path filterExplanation:The `MAX_ACTIVATING_TOKENS` primarily contain "File" and "path" when considered with context. The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` shows "Filter" and "path" appearing after "File". The `TOP_ACTIVATING_TEXTS` include examples of file paths and file filters (`CLAD Files (*.clad), *.clad`). Therefore, the neuron seems to be associated with file paths and filters.Chosen explanation: "file path filter" (3 words)file path filter

    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_31_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
     merge
    0.28
     visuals
    0.27
     SCIENCE
    0.27
    高考
    0.27
     licensed
    0.26
     Sublime
    0.25
    dataType
    0.25
    掩
    0.25
     الاتحاد
    0.25
    展开
    0.24
    POSITIVE LOGITS
    Watts
    0.30
     studs
    0.30
    clearInterval
    0.30
    𒈥
    0.30
    reon
    0.30
     কথোপ
    0.29
    Mkv
    0.28
     spacers
    0.28
     Hous
    0.28
     kleinere
    0.28
    Activations Density 0.012%

    No Known Activations