© 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
    2. Gemma-3-27B-IT
    3. 37-GEMMASCOPE-2-TRANSCODER-262K
    4. 201190
    Prev
    Next
    INDEX
    Explanations

    `minval=1`, `20-80%`, `at least 5 servings`, `at least 6 gauge`, `at least a 3-second following distance`, `sets of 60 seconds`The neuron seems to be detecting numerical values, especially those related to minimums, ranges, or quantities that are part of a specific item or instruction.- `MAX_ACTIVATING_TOKENS` includes `3`, `least`.- `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` shows numbers following specific tokens: `20-80%` contains `20`, `80`.The neuron detects numerical ranges and minimums

    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_37_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
     moderation
    1.13
    Moder
    1.07
    moder
    0.98
     Moder
    0.95
     moder
    0.95
     moderators
    0.88
     Moderator
    0.87
     moderator
    0.85
     Moderators
    0.84
     moderated
    0.82
    POSITIVE LOGITS
     KZ
    0.37
     Large
    0.37
    вото
    0.37
     requisito
    0.36
    弭
    0.36
    ೇವೆ
    0.36
     Stephenson
    0.35
    列
    0.35
     Sexuality
    0.34
    																													
    0.34
    Activations Density 0.035%

    No Known Activations