© 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. 15773
    Prev
    Next
    INDEX
    Explanations

    pm, company, learning, number

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

    The neuron is primarily detecting tokens that occur at the start of new sentences or blocks (i.e. sentence-initial tokens).

    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
     inventory
    0.77
    RE
    0.77
    rived
    0.76
     battlefield
    0.75
    inence
    0.73
    3
    0.73
    9
    0.72
    OURCE
    0.71
    G
    0.71
     ज
    0.71
    POSITIVE LOGITS
    dır
    0.91
    还将
    0.79
    னையும்
    0.76
    将在
    0.75
     době
    0.73
    せい
    0.73
    原生
    0.73
    də
    0.71
    的には
    0.70
     sensitized
    0.70
    Activations Density 0.000%

    No Known Activations