© 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. 11976
    Prev
    Next
    INDEX
    Explanations

    code and technical terms

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

    The neuron responds strongly to rare or unusual tokens—that is, low‐frequency words such as technical terms, names, or infrequently seen vocabulary.

    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
    م
    1.10
    ر
    0.92
    مو
    0.84
    ور
    0.82
    ありません
    0.80
    اری
    0.79
    j
    0.79
    U
    0.77
    Cs
    0.76
    èse
    0.75
    POSITIVE LOGITS
    귿
    0.78
    }}>
    0.76
     그리고
    0.73
    晉
    0.73
    
    0.72
     регистрации
    0.68
    ⇵
    0.66
     প্রেসক্লা
    0.66
    clientID
    0.65
    詓
    0.65
    Activations Density 0.001%

    No Known Activations