© 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. 15113
    Prev
    Next
    INDEX
    Explanations

    odds ratios

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

    The neuron fires on capitalized “name” tokens—i.e. proper‐noun headings such as journal abbreviations (e.g. “J. Org Chem,” “Opt. Commun.”), organization names (“Health,” “Safety,” “Administration”), and similar title words (“Therapy”).

    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
    that
    -2.66
    <bos>
    -2.41
    i
    -2.28
    
    -2.28
    蛲
    -2.17
     and
    -2.11
    
    -2.09
    ところで
    -2.06
     you
    -2.05
    }^{
    -2.05
    POSITIVE LOGITS
    ↵
    3.17
    渼
    2.33
    𖣘
    2.11
    金色
    2.09
     cruelly
    2.09
     of
    2.06
     キーホルダー
    2.03
     применения
    2.00
    浤
    2.00
    ⁛
    1.98
    Activations Density 0.168%

    No Known Activations