© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 84584
    Prev
    Next
    INDEX
    Explanations

    searches, Hooch, Noel

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on a specific subword fragment that appears in various proper names and place names (e.g. the “Hoo” BPE chunk).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ൎ
    -2.61
     avec
    -2.38
    龊
    -2.33
     noteworthy
    -2.25
     This
    -2.22
    ة
    -2.22
    剽
    -2.19
    少ない
    -2.17
     Dieses
    -2.14
    舂
    -2.05
    POSITIVE LOGITS
    他们的
    2.84
    栳
    2.58
    峮
    2.36
          
    2.33
    躥
    2.23
    ми
    2.20
    ↵
    2.20
    完美的
    2.16
    ن
    2.14
     vixion
    2.13
    Activations Density 0.010%

    No Known Activations