© 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. 34-GEMMASCOPE-RES-131K
    4. 109393
    Prev
    Next
    INDEX
    Explanations

    word roots followed by common suffixes

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

    The neuron is sensitive to uncommon, longer subword fragments (i.e. rare or technical character sequences) rather than ordinary short tokens.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    ちょっと
    -1.00
    本当に
    -1.00
    ņš
    -0.97
    ächter
    -0.94
    相続
    -0.94
    ที่
    -0.93
    Côte
    -0.89
     очень
    -0.88
     unleashed
    -0.88
    quite
    -0.87
    POSITIVE LOGITS
     разнообраз
    1.06
    sworth
    0.93
    entious
    0.90
    了一些
    0.88
     Outros
    0.88
    YTICAL
    0.88
    rictive
    0.86
    picious
    0.85
    Precautionary
    0.85
    inerario
    0.85
    Activations Density 0.091%

    No Known Activations