© 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. 7485
    Prev
    Next
    INDEX
    Explanations

    chain of thought and reasoning

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

    The neuron fires strongly on structural metadata markers—e.g. colons in timestamps or headings, navigation breadcrumbs, and citation‐style headings—i.e. it detects document metadata separators.

    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
    debug
    0.86
    forRoot
    0.80
    ]="
    0.77
    aires
    0.77
    致力于
    0.77
    cores
    0.74
    ef
    0.74
    拱
    0.74
     totale
    0.73
    ænd
    0.73
    POSITIVE LOGITS
     قد
    0.88
     appalling
    0.82
     excruciating
    0.81
     حسين
    0.80
     Та
    0.79
     อาจ
    0.79
     ตาม
    0.79
     implications
    0.78
     Некоторые
    0.78
     Кроме
    0.77
    Activations Density 0.000%

    No Known Activations