© 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. 1865
    Prev
    Next
    INDEX
    Explanations

    positional or ordinal number suffixes

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

    The neuron fires strongly on numeric tokens (dates, measurements, counts, etc.)—in other words, it detects numbers 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
    şte
    0.75
     Raffle
    0.74
    ắn
    0.72
    渔
    0.71
    仇
    0.71
    卖家
    0.71
    𝑠
    0.70
     Defin
    0.68
    locale
    0.67
    markt
    0.66
    POSITIVE LOGITS
     ability
    0.75
    AsyncTask
    0.73
     sequenced
    0.71
     eluted
    0.71
     смог
    0.68
     ഒന്നും
    0.68
    I
    0.67
     করিত
    0.67
    Silence
    0.67
     rewritten
    0.66
    Activations Density 0.000%

    No Known Activations