© 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. 2555
    Prev
    Next
    INDEX
    Explanations

    bolding and table separators

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

    The neuron primarily detects emotive icons—both ASCII emoticons (like “:)” or “LOL”) and Unicode emojis—used to express emotion.

    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
     maatau
    0.46
    成员
    0.43
    CameraOpened
    0.42
    埓
    0.42
    厢
    0.42
    Miscellaneous
    0.41
    Acknowledgment
    0.41
    መረ
    0.41
    AlignedText
    0.41
    ൊരു
    0.40
    POSITIVE LOGITS
     Aunque
    0.64
     (
    0.61
     ؟
    0.58
     (?)
    0.57
    ؟
    0.57
    ؟؟
    0.57
     มัน
    0.56
     וא
    0.56
     ;
    0.55
     เพราะ
    0.55
    Activations Density 0.130%

    No Known Activations