© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 12938
    Prev
    Next
    INDEX
    Explanations

    The neuron fires strongly on words and forms that describe physical wearing or erosion (e.g. wear, wear-away, worn).

    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
     watery
    0.66
    关闭
    0.64
     popup
    0.63
     stagnant
    0.62
     pop
    0.60
     impenetrable
    0.59
    呛
    0.59
     complica
    0.59
     stagn
    0.59
    🤧
    0.58
    POSITIVE LOGITS
     scratches
    1.12
     Damage
    1.01
     scratched
    0.99
     Wear
    0.98
     wear
    0.97
     damage
    0.96
     worn
    0.96
    磨
    0.94
     abrasion
    0.93
     scratch
    0.91
    Activations Density 0.270%

    No Known Activations