© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 12967
    Prev
    Next
    INDEX
    Explanations

    license agreement URLs

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

    The neuron fires on key content nouns—especially technical or instructional terms (e.g. tips, wiring, diagram, appearance)—that signal the main subject of a how-to or advice passage.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    words related to do-it-yourself instructions and practical procedures.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    abstract, formal nouns that denote processes, attributes, or components in instructional or technical contexts

    oai_token-act-pair · gpt-5Triggered 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
     sensual
    0.63
     jaundice
    0.63
     January
    0.63
     Miami
    0.59
     Zumba
    0.58
     bisschen
    0.57
     
    0.57
     UCLA
    0.56
     tháng
    0.55
     passionately
    0.55
    POSITIVE LOGITS
    ivad
    0.53
    jš
    0.52
    iby
    0.48
    šu
    0.46
    而不
    0.46
    組立
    0.46
    ául
    0.45
    ješ
    0.44
    isci
    0.44
    openg
    0.44
    Activations Density 0.011%

    No Known Activations