© 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. 22-GEMMASCOPE-RES-131K
    7. 69078
    Prev
    Next
    INDEX
    Explanations

    connecting phrases or lists

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

    The neuron activates on the head noun or key technical term in short two‐word compound phrases (e.g. “international entertainer,” “marginal zone,” “page up,” “J/ψ,” “h_c”).

    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
    デニム
    -1.17
     pâ
    -1.04
     statunit
    -1.03
    诅
    -1.00
    lový
    -0.96
    游戏亮点
    -0.94
     Paroles
    -0.91
     poil
    -0.91
     برخی
    -0.91
     œ
    -0.90
    POSITIVE LOGITS
     and
    1.16
     αλλά
    0.98
    Kjelder
    0.98
     ||
    0.98
    Élet
    0.97
     or
    0.93
     after
    0.92
    뒀
    0.92
    Ṛ
    0.88
     today
    0.88
    Activations Density 0.073%

    No Known Activations