© 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. 7385
    Prev
    Next
    INDEX
    Explanations

    inner layer

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

    The neuron fires on numeric tokens—especially numbers (integers or decimals) and measurement‐style values 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
     combing
    0.74
     fibers
    0.68
     hairs
    0.67
     fémin
    0.65
    鱉
    0.64
     зимо
    0.64
     sleepers
    0.63
    𝐠
    0.63
     fibres
    0.63
     suture
    0.63
    POSITIVE LOGITS
     ก็
    0.85
     hingegen
    0.85
     tiež
    0.81
    の場合は
    0.81
     também
    0.79
     тоже
    0.79
     likewise
    0.78
     tambien
    0.78
     또한
    0.77
     ()
    0.75
    Activations Density 1.163%

    No Known Activations