© 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. 12395
    Prev
    Next
    INDEX
    Explanations

    and or

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

    The neuron fires on uncommon or unusually specific words—that is, it picks out rare or low-frequency vocabulary.

    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
     människor
    1.14
     lidí
    1.07
     whatnot
    1.05
    <unused1728>
    1.02
     município
    0.99
    <unused1879>
    0.99
     ఇతర
    0.98
    <unused589>
    0.95
     paljon
    0.93
    <unused331>
    0.93
    POSITIVE LOGITS
     (=
    1.18
     and
    1.14
     (
    1.13
     =
    1.10
     này
    1.01
     +
    1.00
    й
    1.00
     и
    0.99
    (
    0.99
    ()
    0.98
    Activations Density 0.000%

    No Known Activations