© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10036
    Prev
    Next
    INDEX
    Explanations

    be ridiculous

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

    The neuron is strongly activated by words inside or immediately around quoted dialogue—i.e. it detects direct speech and speaker attributions.

    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
     macho
    0.79
     allegedly
    0.71
     awesome
    0.70
     potholes
    0.66
     unleashing
    0.65
     ingested
    0.64
     paranoid
    0.64
     eyeb
    0.64
     intimid
    0.64
     tường
    0.64
    POSITIVE LOGITS
     Novels
    0.84
    ComboBox
    0.79
    小説
    0.77
    Honestly
    0.75
    Invalids
    0.75
     novels
    0.75
     Settlements
    0.74
    ствуйте
    0.72
    Actors
    0.70
    📘
    0.70
    Activations Density 0.117%

    No Known Activations