© 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. 22-GEMMASCOPE-RES-131K
    7. 64804
    Prev
    Next
    INDEX
    Explanations

    names, places, and brands

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

    The neuron is highly sensitive to short subword tokens (especially three‐ to four‐letter prefixes or suffixes) rather than whole words.

    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
     stakeholders
    -0.95
     می‌گو
    -0.87
    ᴢ
    -0.86
     아니라
    -0.85
    şiv
    -0.84
     проведение
    -0.83
    ₖ
    -0.83
     nového
    -0.83
    ñadir
    -0.82
     그
    -0.81
    POSITIVE LOGITS
    robert
    0.90
     kompakt
    0.86
    ViewChild
    0.86
    來
    0.85
     vacanze
    0.85
    ingat
    0.85
     jedem
    0.84
     respekt
    0.83
     similar
    0.82
    netflix
    0.82
    Activations Density 0.054%

    No Known Activations