© 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. 11
    Prev
    Next
    INDEX
    Explanations

    more, very, extremely

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

    This neuron primarily activates on auxiliary and modal verbs (e.g. “had,” “can,” “be,” “did”) in text.

    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
     mediante
    0.35
     which
    0.35
     atau
    0.33
     -
    0.33
    -
    0.33
     or
    0.32
     dioxide
    0.31
     yoki
    0.31
    0.31
     ,
    0.31
    POSITIVE LOGITS
     खूप
    0.62
     väldigt
    0.59
     veldig
    0.58
    すごく
    0.57
     খুবই
    0.56
     बेहद
    0.55
     엄청
    0.54
     상당히
    0.54
     काफी
    0.54
     굉장히
    0.54
    Activations Density 15.536%

    No Known Activations