© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 10464
    Prev
    Next
    INDEX
    Explanations

    Final Answer

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

    The neuron fires strongly on lines of metadata that list categories, tags, or sections (e.g. “posted in … and tagged …”), effectively detecting blog‐style category/tag headers.

    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
    ).
    0.68
     detik
    0.66
    )
    0.66
     nóng
    0.65
     hopelessness
    0.63
    owych
    0.61
     foil
    0.61
    ),
    0.61
     spoonful
    0.61
     shone
    0.60
    POSITIVE LOGITS
    ließlich
    0.73
    Giz
    0.72
     Algunos
    0.71
     Außerdem
    0.69
    ेक्षा
    0.67
     Agricult
    0.65
    プリング
    0.64
    瓤
    0.64
     Serm
    0.63
     Microbial
    0.63
    Activations Density 0.894%

    No Known Activations