© 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. 38-GEMMASCOPE-2-TRANSCODER-262K
    4. 1366
    Prev
    Next
    INDEX
    Explanations

    The neuron detects date formats and separators, often associated with structured text or separators. It seems to be recognizing common delimiters and numerical patterns found in dates and data entries.Looking at the prompt, the MAX_ACTIVATING_TOKENS contains only '/', '-', and '/'. These are common separators.The TOKENS_AFTER_MAX_ACTIVATING_TOKEN contains '0', 'sep', '0', '1', '1', '0', '0', '0', '0', '0'. 'sep' is a strong indicator. The numbers '0' and '1' might also relate to flags or counts.date delimiters or separators

    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_38_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
     frictional
    0.38
    oths
    0.37
     cai
    0.36
     περισ
    0.35
     ক্ষুদ্র
    0.35
     sert
    0.34
     왔다
    0.34
    غیر
    0.33
     tel
    0.33
     vita
    0.33
    POSITIVE LOGITS
     Gaye
    0.47
    "][
    0.38
     মাহমুদ
    0.37
    "]:
    0.37
     Zas
    0.36
    endus
    0.35
    "]))
    0.35
    ]}/${
    0.35
    सही
    0.34
    〽
    0.34
    Activations Density 0.011%

    No Known Activations