© 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. 11281
    Prev
    Next
    INDEX
    Explanations

    same or again

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

    The neuron fires on formal “boilerplate” cue words that introduce or qualify clauses—especially quantifiers and provisos like “includes,” “all,” “any,” “provided,” “for,” and “from.”

    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
    下の
    0.62
    για
    0.54
    www
    0.52
    H
    0.52
    ೪
    0.51
    שים
    0.51
    事前に
    0.50
     H
    0.50
    define
    0.50
    深刻
    0.50
    POSITIVE LOGITS
    同样的
    2.27
    同样
    2.23
    同樣
    2.20
     dezelfde
    2.06
     again
    2.05
     ebenfalls
    2.03
     novamente
    2.00
    同様
    1.99
     hetzelfde
    1.98
     nuevamente
    1.95
    Activations Density 1.333%

    No Known Activations