© 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. 12504
    Prev
    Next
    INDEX
    Explanations

    talking about challenge

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

    The neuron activates on occurrences of the word “challenge” (and its morphological variants such as “challenging”).

    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
     مساله
    0.86
     impediments
    0.86
    の問題
    0.72
     hurdles
    0.68
     obstacles
    0.67
     hindrance
    0.65
     *=
    0.64
     hardships
    0.64
     চাপ
    0.63
     impediment
    0.63
    POSITIVE LOGITS
    ACCEPTED
    0.77
    de
    0.73
    ENGE
    0.70
    ोटी
    0.66
    et
    0.66
     kunt
    0.65
    achricht
    0.65
    ాగ
    0.64
    ิ
    0.64
    ৬
    0.64
    Activations Density 0.080%

    No Known Activations