© 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. 73375
    Prev
    Next
    INDEX
    Explanations

    phrases of composition or relation

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

    The neuron fires strongly on relational function words (especially prepositions and conjunctions) that link entities—e.g. “of,” “from,” “or.”

    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
     cardiovas
    -1.16
    Сере
    -1.10
    Spesifikasi
    -1.10
    よくある
    -1.09
     بسبب
    -1.08
     ecos
    -1.08
     cardia
    -1.08
     sorri
    -1.08
    最近は
    -1.07
    ユニクロ
    -1.07
    POSITIVE LOGITS
    (),
    1.09
    !,
    1.02
     at
    0.95
     mer
    0.95
    García
    0.94
     [],
    0.93
     one
    0.92
     Work
    0.91
     in
    0.91
     by
    0.90
    Activations Density 0.169%

    No Known Activations