© 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. 10419
    Prev
    Next
    INDEX
    Explanations

    greeting prompts

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

    The neuron fires on unusually rare or “out‐of‐vocabulary” tokens—things like citation codes, page ranges, uncommon nouns, or other oddball tokens that barely ever occur in normal 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
     entanto
    0.96
    可是
    0.80
    ibid
    0.80
    е
    0.79
     vlá
    0.75
    ὶ
    0.75
    о
    0.73
     государ
    0.73
    ud
    0.72
     mì
    0.72
    POSITIVE LOGITS
    BANG
    0.79
    🪁
    0.79
    Sco
    0.78
    میری
    0.77
    𝙎
    0.76
    Notre
    0.73
    bly
    0.73
    SNAP
    0.73
    🟢
    0.73
     Mojo
    0.73
    Activations Density 0.000%

    No Known Activations