© 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. 34-GEMMASCOPE-RES-131K
    7. 57343
    Prev
    Next
    INDEX
    Explanations

    avoid or reduce

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

    The neuron strongly fires on words that appear in medical/disclaimer contexts—especially warning or advisory terms and conjunctions (e.g. “or,” “becomes,” “reduce,” “disregard,” “failure”).

    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
     finalement
    -0.82
     masih
    -0.82
    やら
    -0.78
    وأضاف
    -0.77
    Bises
    -0.76
    ibilidad
    -0.75
    eckt
    -0.74
    sequently
    -0.73
    Outro
    -0.73
     защиту
    -0.72
    POSITIVE LOGITS
     pause
    0.91
     relocated
    0.83
    archiv
    0.82
    ziesz
    0.82
     sideways
    0.80
     suspended
    0.79
     severely
    0.78
    getMin
    0.77
     reduced
    0.77
     millennia
    0.75
    Activations Density 0.076%

    No Known Activations