© 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. 31-GEMMASCOPE-2-RES-16K
    7. 2126
    Prev
    Next
    INDEX
    Explanations

    numbers or letters

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

    The neuron fires strongly on long alphanumeric “ID” or code‐like tokens (strings of letters and digits), effectively spotting encoded identifiers rather than ordinary words.

    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_31_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
     reembolso
    0.89
     aprender
    0.84
     १८४
    0.78
     õ
    0.74
     entretenimiento
    0.74
     aceite
    0.73
     pérdida
    0.73
     १००
    0.72
     évidemment
    0.71
     надеюсь
    0.71
    POSITIVE LOGITS
    Q
    0.90
    Container
    0.88
    Slice
    0.88
    EventEmitter
    0.88
    Con
    0.88
    Effect
    0.88
    Field
    0.85
    AA
    0.85
    Var
    0.85
    L
    0.85
    Activations Density 0.033%

    No Known Activations