© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 14628
    Prev
    Next
    INDEX
    Explanations

    support, cost, point, and inferior

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

    The neuron is effectively acting as a positional marker, firing most strongly on tokens that occur near the very beginning of a new document segment.

    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
    る
    0.80
    Drum
    0.79
    Chevrolet
    0.77
    Nissan
    0.77
    AIS
    0.75
    GLUT
    0.75
    DIR
    0.74
    Arch
    0.74
    ím
    0.73
    impossible
    0.73
    POSITIVE LOGITS
    itories
    0.78
    ແມ່ນ
    0.75
    شاہ
    0.73
     санк
    0.73
    人士
    0.72
    _{*}(
    0.72
    ۥ
    0.72
     источники
    0.72
    ders
    0.71
     deems
    0.71
    Activations Density 0.000%

    No Known Activations