© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 2031
    Prev
    Next
    INDEX
    Explanations

    error messages with colons

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

    The neuron fires on technical identifiers or labels—i.e. code‐ or config‐style field names, keys, and other uppercase or CamelCase tokens.

    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
    People
    0.55
    自分が
    0.51
    Map
    0.48
    Holiday
    0.46
    ‘
    0.45
    Get
    0.44
    Info
    0.44
    Phone
    0.44
    Fleet
    0.44
    SOME
    0.44
    POSITIVE LOGITS
    :",
    1.44
    :");
    1.30
     :");
    1.27
    .",
    1.22
    :")
    1.21
    !",
    1.15
     :',
    1.14
    !");
    1.13
    .");
    1.11
    .")
    1.09
    Activations Density 0.030%

    No Known Activations