© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5539
    Prev
    Next
    INDEX
    Explanations

    The neuron simply responds to the length of a token, activating strongly for longer words regardless of their meaning.

    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_24_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
    ي
    0.98
     כא
    0.93
     моей
    0.93
    আমার
    0.91
     мире
    0.91
     подар
    0.88
     откро
    0.87
     ө
    0.85
     আমার
    0.85
     этом
    0.85
    POSITIVE LOGITS
    oL
    0.80
    folio
    0.75
    PR
    0.70
    offsetHeight
    0.70
     compass
    0.70
    PA
    0.68
    PER
    0.67
    Port
    0.66
    Point
    0.66
    OB
    0.66
    Activations Density 0.000%

    No Known Activations