© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 104548
    Prev
    Next
    INDEX
    Explanations

    each and every

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

    The neuron fires on the word “Each,” marking when a sentence begins with that quantifier (e.g. “Each week,” “Each year,” etc.).

    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
    </strong>
    -2.27
    -
    -2.14
    ar
    -2.05
    H
    -1.99
     what
    -1.94
     who
    -1.75
    }
    -1.63
    us
    -1.59
    党的
    -1.56
    个
    -1.55
    POSITIVE LOGITS
    ࿚
    2.02
    rália
    1.93
    pidou
    1.91
     abſ
    1.87
    geoning
    1.82
    timento
    1.82
    苳
    1.80
    谖
    1.79
    遹
    1.78
    ribue
    1.77
    Activations Density 0.017%

    No Known Activations