© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 994
    Prev
    Next
    INDEX
    Explanations

    names followed by titles or roles

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find numeric and statistical tokens—especially digits, timestamps, scores, and other number-heavy data.

    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_31_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
    Пі
    0.52
    END
    0.52
    Бо
    0.52
    ПА
    0.52
    У
    0.51
    Б
    0.51
    ATTR
    0.50
    𝟕
    0.50
    CONNECTION
    0.49
    Бу
    0.49
    POSITIVE LOGITS
     sogenannten
    0.66
     empresa
    0.65
     perusahaan
    0.63
     Utilis
    0.61
     pihak
    0.58
     Namun
    0.57
     Geschäftsführer
    0.57
     Jenis
    0.57
     Mga
    0.56
     pemain
    0.56
    Activations Density 0.339%

    No Known Activations