© 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. 3474
    Prev
    Next
    INDEX
    Explanations

    explaining growth

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

    This neuron strongly activates on tokens that are specific named‐entity snippets (initials, proper names) and numeric fragments (years, percentages, digits).

    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
     isomorphic
    0.73
     тип
    0.71
    тип
    0.67
    ෲ
    0.67
    ೬
    0.64
     sinon
    0.63
     carcinomas
    0.63
     subcl
    0.63
     digunakan
    0.62
     С
    0.61
    POSITIVE LOGITS
     tonight
    0.90
     Tuesday
    0.89
     이날
    0.87
     afgelopen
    0.85
     Tonight
    0.85
    Tuesday
    0.84
     Wednesday
    0.83
    ursday
    0.83
    aturday
    0.82
     Thursday
    0.82
    Activations Density 0.096%

    No Known Activations