© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 109049
    Prev
    Next
    INDEX
    Explanations

    bad source, checked, heart failure, moonrise, real object, true, 9

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

    The neuron fires on standalone capitalized keywords that label sections or statuses (e.g. “Bad,” “Good,” “checked,” “Real,” “True”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Transcriptie
    -0.86
    Transkript
    -0.84
    ceny
    -0.82
    ставки
    -0.82
    ūnas
    -0.79
     akrab
    -0.78
     suiker
    -0.77
    Pry
    -0.77
    ！】
    -0.77
    impin
    -0.76
    POSITIVE LOGITS
    especially
    0.87
    Especially
    0.84
    leaving
    0.81
    Is
    0.79
    Якщо
    0.79
     Bescheid
    0.78
     Всё
    0.78
     Национальный
    0.77
    置信
    0.77
    @[+][
    0.76
    Activations Density 0.001%

    No Known Activations