© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 14293
    Prev
    Next
    INDEX
    Explanations

    Python code structure, module, and work

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

    The neuron is broadly looking for mid-frequency content and function words—tokens (including both some pronouns like “you/your/not” and longer content terms such as “cookies,” “coordinating,” “setting,” or “inadvertently”) that occur neither extremely often nor extremely rarely.

    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
    <0x0D>
    0.85
    )</
    0.82
    )$
    0.74
     commer
    0.73
    ↵
    0.73
    )=\
    0.71
     <?
    0.71
     gd
    0.71
     gn
    0.71
    ).
    0.70
    POSITIVE LOGITS
    に基づく
    0.93
    テナンス
    0.91
    プター
    0.90
    ギリス
    0.88
     sobrevivir
    0.88
     aHUS
    0.87
    ewnątrz
    0.85
     выяв
    0.84
    姆
    0.84
     dioxygen
    0.83
    Activations Density 0.004%

    No Known Activations