© 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. 22-GEMMASCOPE-RES-131K
    7. 26505
    Prev
    Next
    INDEX
    Explanations

    results and outcomes

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

    The neuron responds strongly to very common/frequent tokens—especially the word “and” and standalone numbers—i.e. it’s detecting high-frequency function words and numeric tokens.

    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
    iteur
    -0.85
     மற்றும்
    -0.84
    ukas
    -0.81
    "}")
    -0.80
    Osob
    -0.80
     ст
    -0.80
    κο
    -0.79
     If
    -0.79
     bereiken
    -0.79
    HTTPResponse
    -0.77
    POSITIVE LOGITS
     zero
    0.86
     scoreless
    0.85
    ียญ
    0.80
     condam
    0.77
     نگ
    0.76
    ꧁
    0.75
     gefe
    0.75
    Tun
    0.73
     wherein
    0.73
    declare
    0.73
    Activations Density 0.159%

    No Known Activations