© 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. 5339
    Prev
    Next
    INDEX
    Explanations

    coding, technical terms, names

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

    The neuron fires on words that signal a shift into the author’s personal commentary or side‐note (e.g. conditional “if,” personal “you,” gerund “trying,” bracketed citation “[Internet],” negation “didn’t,” descriptive “fluffy,” etc.).

    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
    bake
    0.77
    dims
    0.75
     iż
    0.74
     previstas
    0.73
    mui
    0.72
    biz
    0.70
    օ
    0.70
     ksi
    0.70
    ból
    0.69
     huts
    0.69
    POSITIVE LOGITS
    ر
    0.85
    不存在
    0.79
    עות
    0.78
    登录
    0.76
     biblioteca
    0.75
     soluzione
    0.75
    Número
    0.75
     通知
    0.75
    verifica
    0.74
     食
    0.74
    Activations Density 0.000%

    No Known Activations