© 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. 10-GEMMASCOPE-RES-131K
    7. 38949
    Prev
    Next
    INDEX
    Explanations

    formación

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

    This neuron fires most strongly on isolated “term‐label” tokens—like section headings or vocabulary labels (e.g. “розділ”/“ділу,” “термін,” “транслитерації,” “Sixth Form,” “остальные”)—i.e. words that name or label document sections or key terms.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    1
    -3.11
    5
    -3.03
    7
    -2.73
    ↵↵
    -2.66
    4
    -2.66
    0
    -2.59
    -
    -2.44
    6
    -2.33
     shove
    -2.27
    ～、
    -2.25
    POSITIVE LOGITS
    呖
    2.86
    櫪
    2.73
    辂
    2.70
    лі
    2.66
    蹰
    2.63
    麁
    2.63
    窳
    2.58
    蹋
    2.58
     nuovi
    2.55
     need
    2.53
    Activations Density 0.002%

    No Known Activations