© 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. 22-GEMMASCOPE-RES-131K
    7. 75055
    Prev
    Next
    INDEX
    Explanations

    numbers and quantities

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

    The neuron responds to infrequent or content‐specific tokens—especially numbers, uncommon technical terms, and other low‐frequency words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     twar
    -0.94
     зеленый
    -0.90
    𝒾
    -0.89
     Trong
    -0.87
    階級
    -0.86
     =-\
    -0.86
     arbejds
    -0.86
    𝘌
    -0.86
     Tit
    -0.85
    .")
    -0.84
    POSITIVE LOGITS
    ්‍ර
    0.86
     figlio
    0.85
     fünf
    0.81
     duiz
    0.78
    nier
    0.78
    夕食
    0.77
     sechs
    0.75
    sela
    0.75
    paginate
    0.75
     falta
    0.74
    Activations Density 0.008%

    No Known Activations