© 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. 24-GEMMASCOPE-2-RES-16K
    7. 9853
    Prev
    Next
    INDEX
    Explanations

    files with documentation or context

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

    The neuron fires most strongly on semantically rich, content-bearing words—longer nouns, verbs, or adjectives (e.g. “pleasure,” “hide,” “pain,” “discontentedly,” “countries,” “plans”) rather than on function words.

    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_24_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
     দেরী
    0.78
     reinst
    0.68
     inject
    0.66
     replen
    0.65
    ようになった
    0.65
    っかり
    0.64
    𝗔
    0.64
     여기에
    0.64
     extinguish
    0.63
     oscillation
    0.62
    POSITIVE LOGITS
    ad
    1.05
    bild
    1.04
    onk
    1.00
    anak
    0.97
    on
    0.96
    onov
    0.96
    ay
    0.95
    luğu
    0.95
    ר
    0.93
    onos
    0.92
    Activations Density 0.000%

    No Known Activations