© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10961
    Prev
    Next
    INDEX
    Explanations

    substitute for advice

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

    The neuron strongly activates on the key action words and content terms used in procedural or “how-to” instructions.

    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
    Hydrochloride
    0.89
    zneim
    0.87
    ائص
    0.84
    ྪ
    0.84
    ítése
    0.84
    extrémité
    0.83
     Calyce
    0.83
    umlahan
    0.82
    impulan
    0.82
    itochond
    0.80
    POSITIVE LOGITS
     I
    0.96
    0.94
     one
    0.92
     later
    0.92
     two
    0.90
     day
    0.90
     already
    0.87
     more
    0.86
     the
    0.85
     not
    0.84
    Activations Density 0.025%

    No Known Activations