© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 3940
    Prev
    Next
    INDEX
    Explanations

    `&` and `game`

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

    The neuron fires strongly on the structural “metadata” of forum‐style posts—things like the bits of text around quoted passages (apostrophes, colons, user‐name markers, “wrote:” lines, section headers, post IDs, timestamps, numeric markers, 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
     здійсню
    0.70
     omnichannel
    0.69
     inextricably
    0.69
    ුණු
    0.69
     здійс
    0.68
     transformative
    0.67
    ⾃
    0.66
     thereby
    0.66
     revital
    0.64
     philanthrop
    0.64
    POSITIVE LOGITS
    /
    1.00
     tengo
    0.95
     pics
    0.91
     ebay
    0.91
    +
    0.89
     Tapatalk
    0.89
     tbh
    0.86
     bbq
    0.86
     crappy
    0.85
    mmm
    0.83
    Activations Density 0.302%

    No Known Activations