© 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. 10305
    Prev
    Next
    INDEX
    Explanations

    bot, database, risk, contra, debt, front

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

    The neuron fires strongly on formatting‐style tokens (backslashes “\” and runs of underscores such as “____”) that occur in category paths or fill-in-the-blank placeholders.

    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
     figli
    0.82
    㭉
    0.80
    Plaintiffs
    0.79
    뜩
    0.79
    γά
    0.78
    ας
    0.77
     collapsing
    0.76
    ם
    0.76
     екс
    0.75
     contenant
    0.75
    POSITIVE LOGITS
    u
    0.91
    ur
    0.82
    at
    0.82
    ing
    0.75
    už
    0.74
    al
    0.72
    ystone
    0.71
    il
    0.71
     sports
    0.71
    orative
    0.70
    Activations Density 0.000%

    No Known Activations