© 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-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 21893
    Prev
    Next
    INDEX
    Explanations

    adverbs and related derivatives

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

    This neuron activates on common 5-letter character sequences that form the beginnings of longer words (e.g. “gener,” “tempor,” “perma,” “compar,” “exclu,” etc.).

    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
     mutta
    -3.31
    趼
    -2.89
     But
    -2.55
    慊
    -2.53
     tetapi
    -2.52
    ↑↑↑</
    -2.44
    خدام
    -2.42
    窠
    -2.39
    🫔
    -2.39
    也很
    -2.38
    POSITIVE LOGITS
    {
    3.22
    ."
    3.11
    鼉
    2.81
     这
    2.61
    ي
    2.59
    ............
    2.58
    ly
    2.56
    ("
    2.55
     craz
    2.53
    .............
    2.50
    Activations Density 0.075%

    No Known Activations