© 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. 6804
    Prev
    Next
    INDEX
    Explanations

    chain

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

    The neuron selectively fires for longer or unusual tokens—especially multi‐syllabic words, proper names, or other low‐frequency terms.

    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
    Machines
    0.66
    arci
    0.66
    고
    0.65
    凑
    0.64
    zij
    0.62
    ated
    0.62
    й
    0.61
    hlen
    0.61
    켤
    0.61
    ်
    0.60
    POSITIVE LOGITS
    tenance
    0.83
    テナンス
    0.75
     multitudes
    0.73
    tained
    0.71
    つまり
    0.71
    taining
    0.66
    ଁ
    0.66
    tainment
    0.64
    bursement
    0.63
     অভিষ
    0.63
    Activations Density 0.710%

    No Known Activations