© 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. 10-GEMMASCOPE-RES-131K
    7. 39097
    Prev
    Next
    INDEX
    Explanations

    that resulted in

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

    The neuron fires on news‐style “that” clauses—i.e. noun or event phrases immediately followed by “that” plus a verb (e.g. “spree that put…,” “scandal that could…,” “operation that moved…,” “ring that rotated…”).

    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
    Surprisingly
    -1.84
    Often
    -1.80
     there
    -1.77
    Sometimes
    -1.73
    Unlike
    -1.70
    There
    -1.69
    Usually
    -1.66
    While
    -1.66
    ferences
    -1.64
    Knowing
    -1.64
    POSITIVE LOGITS
     econó
    2.02
     exitos
    1.73
     téc
    1.70
     pasillo
    1.70
     ám
    1.66
    mpagne
    1.66
    見えます
    1.63
     két
    1.63
    
    1.60
     campamento
    1.59
    Activations Density 0.062%

    No Known Activations