© 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. 24-GEMMASCOPE-2-RES-16K
    7. 2558
    Prev
    Next
    INDEX
    Explanations

    locations and directions

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

    The neuron detects directional or spatial adverbs/prepositions (e.g. “up,” “down,” “on,” “along,” “both”).

    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
    EOUT
    0.59
    𝖊
    0.57
    forma
    0.57
    OL
    0.55
    EC
    0.55
     आशंका
    0.54
    ক
    0.54
    ባድ
    0.54
    জন্য
    0.54
    ніка
    0.54
    POSITIVE LOGITS
     coasts
    0.91
     beaches
    0.83
     estuaries
    0.76
     sidewalks
    0.75
     weekdays
    0.73
     transients
    0.73
     steroids
    0.72
     rivers
    0.72
     pantai
    0.70
     weekends
    0.69
    Activations Density 0.001%

    No Known Activations