© 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. 45087
    Prev
    Next
    INDEX
    Explanations

    JSONObject, Lenovo, Monday

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

    The neuron fires on tokens that begin with a capital letter—especially code identifiers and class/library names (e.g. JSONArray, JSONObject, ByteBuffer) or proper nouns like “Lenovo.”

    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
    h
    -2.41
    ("'"
    -2.20
     Why
    -2.20
    渲
    -2.16
    begin
    -2.16
    ,"
    -2.13
     to
    -2.11
     indelible
    -2.09
    这种
    -2.06
    壕
    -2.06
    POSITIVE LOGITS
     để
    2.84
    神
    2.78
    媖
    2.64
    很赞
    2.53
    because
    2.42
    鯓
    2.39
    \
    2.31
    숑
    2.27
    年輕
    2.25
    總是
    2.25
    Activations Density 0.001%

    No Known Activations