© 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. 10-GEMMASCOPE-RES-131K
    7. 113385
    Prev
    Next
    INDEX
    Explanations

    finding flaws

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

    the word "flaws" or its variations, particularly when discussing defects, problems, or imperfections in something.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The main thing this neuron does is detect mentions of flaws or defects.

    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
    谖
    -2.89
    戩
    -2.88
    罣
    -2.72
    żeli
    -2.59
     現貨
    -2.53
    鲻
    -2.50
     eſſ
    -2.48
     It
    -2.45
    玧
    -2.39
     relates
    -2.38
    POSITIVE LOGITS
    '
    2.98
    你的
    2.63
    g
    2.53
    在我
    2.33
    i
    2.30
    是啊
    2.28
    ac
    2.28
    x
    2.27
     הן
    2.25
    在他
    2.23
    Activations Density 0.002%

    No Known Activations