© 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. 34-GEMMASCOPE-RES-131K
    7. 92273
    Prev
    Next
    INDEX
    Explanations

    upside down, inside out

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

    The neuron activates on compound orientation-reversal phrases (e.g. “upside down” or “inside-out”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    歧
    -0.84
    VERTICAL
    -0.83
     eyeball
    -0.81
    Vertical
    -0.77
    IsDeleted
    -0.77
    Thời
    -0.77
    leti
    -0.75
     Spital
    -0.74
    phils
    -0.74
    itian
    -0.73
    POSITIVE LOGITS
     down
    0.94
     out
    0.82
     вниз
    0.79
    铂
    0.71
     xuống
    0.70
    lname
    0.69
    Digite
    0.69
     stimulation
    0.69
    ]$}
    0.69
    Down
    0.68
    Activations Density 0.023%

    No Known Activations