© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 4402
    Prev
    Next
    INDEX
    Explanations

    The neuron isn’t keyed to any particular part‐of‐speech or semantic category but rather lights up on relatively rare, content‐bearing tokens (i.e. unusual or informative words) in the text.

    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
    ('
    0.76
     (
    0.76
    *,
    0.73
    (),
    0.71
    。
    0.71
    S
    0.67
    Object
    0.66
    ("
    0.66
    （
    0.65
    <b>
    0.64
    POSITIVE LOGITS
     œufs
    0.88
     tengo
    0.83
     ponemos
    0.83
     seguimos
    0.79
     hacemos
    0.78
     université
    0.77
     estamos
    0.76
     huile
    0.75
     Colombeau
    0.75
     फटाफट
    0.74
    Activations Density 0.001%

    No Known Activations