© 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. 22-GEMMASCOPE-RES-131K
    4. 74934
    Prev
    Next
    INDEX
    Explanations

    "to" prepositions

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

    The neuron is picking out long, uncommon or highly technical domain‐specific tokens (e.g. specialized jargon or code identifiers).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     ネイビー
    -0.98
     FRAMEWORK
    -0.90
    UrlEncoded
    -0.88
     exportación
    -0.88
     remarquer
    -0.88
     precisam
    -0.86
     เข
    -0.84
    edì
    -0.82
     Freeze
    -0.81
    aimer
    -0.81
    POSITIVE LOGITS
     могло
    0.95
    Sä
    0.90
     oddly
    0.85
    telijke
    0.84
     могли
    0.84
     menti
    0.83
     pilihan
    0.83
     bereits
    0.82
    fällt
    0.81
    ২
    0.81
    Activations Density 0.006%

    No Known Activations