© 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. 15814
    Prev
    Next
    INDEX
    Explanations

    single, pairwise, required, gain, take

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

    The neuron fires primarily on content-bearing verbs (especially reporting or action verbs) in running text.

    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
    我知道
    0.85
    ぁ
    0.85
    んですが
    0.83
    我
    0.83
    Yea
    0.80
    ヾ
    0.78
    which
    0.76
    哎
    0.76
     ডিসেম্বর
    0.75
    Feb
    0.74
    POSITIVE LOGITS
    ဋ
    0.85
    মাণ
    0.77
    oč
    0.77
    nál
    0.76
     ín
    0.75
    ná
    0.75
    s
    0.75
    uksen
    0.74
    ný
    0.73
    ennium
    0.72
    Activations Density 0.000%

    No Known Activations