© 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. 12060
    Prev
    Next
    INDEX
    Explanations

    persistently stored data

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find relatively rare or low‐frequency content words (i.e. uncommon vocabulary tokens).

    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
     damned
    0.78
    yim
    0.77
    e
    0.75
    ز
    0.74
    of
    0.73
    뵈
    0.73
     of
    0.72
     ซ
    0.70
     damn
    0.69
     mortgage
    0.67
    POSITIVE LOGITS
     respostas
    0.90
    を使用した
    0.81
     demás
    0.80
    を使用して
    0.80
     능력
    0.80
    れる
    0.79
    によって
    0.79
    していた
    0.79
     usare
    0.79
    蝋
    0.79
    Activations Density 0.000%

    No Known Activations