© 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. 54520
    Prev
    Next
    INDEX
    Explanations

    overcome disadvantages or limitations

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

    The neuron activates on words that signal overcoming or resolving problems—e.g. “resolves,” “overcome,” “reduce,” “address,” and similar terms that indicate fixing shortcomings.

    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
    ólico
    -1.12
    haping
    -1.09
    neath
    -1.09
    barten
    -1.08
     ー
    -1.04
     emin
    -1.04
     errone
    -1.04
     disagreements
    -1.03
     orto
    -1.02
    haped
    -1.02
    POSITIVE LOGITS
     often
    1.20
     like
    1.06
     commonly
    1.02
     souvent
    1.01
     currently
    1.00
     vaak
    1.00
     "
    0.98
     spesso
    0.94
     who
    0.93
     часто
    0.93
    Activations Density 0.064%

    No Known Activations