© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 38514
    Prev
    Next
    INDEX
    Explanations

    of [something]

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

    The neuron is highly selective for the preposition “of,” especially in title-style or descriptive phrases (e.g. “city of rich history,” “life of science”).

    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
    着一
    -0.97
    the
    -0.96
     link
    -0.94
    throughout
    -0.94
    atization
    -0.93
     view
    -0.93
    金
    -0.93
     amount
    -0.91
    ii
    -0.90
    着的
    -0.90
    POSITIVE LOGITS
     utmost
    1.59
     contrasts
    1.55
     sorts
    1.52
     considerable
    1.36
     constant
    1.28
     pure
    1.28
     extremes
    1.22
     ideias
    1.21
     which
    1.19
     mostly
    1.17
    Activations Density 0.055%

    No Known Activations