© 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. 3201
    Prev
    Next
    INDEX
    Explanations

    before

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

    This neuron strongly activates on numerical or temporal‐reference tokens (e.g. numbers and the word “Before”).

    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
     bluff
    1.02
     hoarding
    0.93
     kilowatt
    0.89
     Klinefelter
    0.86
    堣
    0.86
    جراء
    0.84
     disenfranch
    0.84
     defer
    0.80
     pebble
    0.80
     dismay
    0.80
    POSITIVE LOGITS
    eminent
    1.09
    ä
    1.03
     الميلاد
    0.99
     وبعد
    0.98
     memulai
    0.93
     tomu
    0.92
    ě
    0.91
     মণ্ডল
    0.91
    hled
    0.88
    emptive
    0.88
    Activations Density 0.059%

    No Known Activations