© 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. 6239
    Prev
    Next
    INDEX
    Explanations

    error messages

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

    The neuron strongly activates on numeric tokens and numeric sequences (e.g. years, decimal numbers, or numbers in dates).

    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
     س
    0.59
    氇
    0.55
    𝙇
    0.52
    规则
    0.51
    จ
    0.51
    ド
    0.51
     ដ
    0.50
    蕖
    0.50
     ج
    0.50
     মো
    0.50
    POSITIVE LOGITS
     payoffs
    0.59
     score
    0.54
     trov
    0.53
     payoff
    0.52
     cheering
    0.52
     foci
    0.52
     oars
    0.52
     alas
    0.52
     sheath
    0.52
     looping
    0.52
    Activations Density 0.004%

    No Known Activations