© 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. 12834
    Prev
    Next
    INDEX
    Explanations

    Singlish conversations

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find words used in procedural or instructional contexts—especially imperative or step‐like tokens (e.g. “open,” “choose,” “send,” “given,” “not”).

    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
    тна
    0.86
     deja
    0.76
    ופ
    0.75
     côte
    0.73
     업체
    0.72
    ・・・。
    0.69
    乇
    0.68
     fréquent
    0.67
    。
    0.66
    دل
    0.63
    POSITIVE LOGITS
    rema
    0.93
    EN
    0.91
     Foo
    0.90
    ir
    0.88
    peripheral
    0.88
    id
    0.86
    plugin
    0.82
    preprocess
    0.79
    robot
    0.78
    error
    0.77
    Activations Density 0.000%

    No Known Activations