© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 38530
    Prev
    Next
    INDEX
    Explanations

    element types

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

    The neuron detects mentions of “elements” (and classical element names like water, fire, earth, air, 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_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
    ząt
    -1.03
    青い
    -0.93
     septembre
    -0.84
    banez
    -0.84
    zaak
    -0.84
    tigste
    -0.83
    Халы
    -0.82
     professionnel
    -0.81
    bigger
    -0.81
    धिक
    -0.80
    POSITIVE LOGITS
     water
    1.67
     element
    1.45
     earth
    1.41
     Earth
    1.34
     elements
    1.30
     стихи
    1.25
    元素
    1.21
     fire
    1.21
     nature
    1.20
     air
    1.16
    Activations Density 0.016%

    No Known Activations