© 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. 10347
    Prev
    Next
    INDEX
    Explanations

    surprisingly complex or essentially describing

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

    The neuron is detecting evaluative/discourse-marker words that signal the author’s attitude or stance (e.g. “unfortunately,” “fortunately,” “probably,” “actually,” etc.).

    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
    cómo
    0.73
    Kako
    0.71
    ҷ
    0.70
    Jeśli
    0.64
     possibilità
    0.63
     compréhension
    0.62
    ্রিয়া
    0.61
    Cómo
    0.61
    éclair
    0.61
     chrét
    0.60
    POSITIVE LOGITS
     (
    0.88
     زیادی
    0.78
     not
    0.75
     (?)
    0.72
     residing
    0.71
     also
    0.70
    也有
    0.70
     coincides
    0.70
    有不少
    0.70
     with
    0.70
    Activations Density 0.406%

    No Known Activations