© 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. 7589
    Prev
    Next
    INDEX
    Explanations

    repeat steps

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

    The neuron is essentially acting as a “high‐magnitude token” detector—it lights up on relatively rare or salient content words with unusually large embedding norms.

    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
    j
    0.92
     a
    0.87
    a
    0.84
    হ
    0.84
    zás
    0.80
    ді
    0.80
    до
    0.79
     delim
    0.79
    r
    0.78
    დ
    0.76
    POSITIVE LOGITS
     semasa
    0.82
    .).
    0.81
     Marbella
    0.81
    .”.
    0.80
     vaik
    0.78
    !”.
    0.77
     سود
    0.77
    !).
    0.74
    ો
    0.73
    Clarke
    0.73
    Activations Density 0.000%

    No Known Activations