© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 13254
    Prev
    Next
    INDEX
    Explanations

    This neuron detects descriptions of vehicle damage and repair, especially references to a scuffed bumper lowering value and being restored.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    jectory
    0.49
    密钥
    0.48
    योग
    0.47
    深入
    0.47
     Midwest
    0.46
    篭
    0.46
     وبركاته
    0.46
    炕
    0.46
    addData
    0.46
     hozzá
    0.46
    POSITIVE LOGITS
     aesthetically
    2.81
     aesthetic
    2.48
     aesthetics
    2.48
     эсте
    2.39
     estética
    2.33
     appearance
    2.23
     Aesthetic
    2.23
     esth
    2.22
    颜值
    2.20
     Aesthetics
    2.20
    Activations Density 0.979%

    No Known Activations