© 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. 11947
    Prev
    Next
    INDEX
    Explanations

    AI models, GANs, Diffusion, Neural Networks

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

    The neuron selectively fires on single, salient content words—mostly nouns or headings like “endings,” “links,” “consultations,” “recourse,” or “speaker”—that introduce or mark a new topic or section.

    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
    क्तिक
    0.79
     machen
    0.73
     angla
    0.71
     alimentos
    0.71
     aliments
    0.70
     धारण
    0.70
    োহণ
    0.69
    g
    0.68
     maken
    0.67
     meine
    0.67
    POSITIVE LOGITS
     REL
    0.82
    Pra
    0.79
     Zaw
    0.75
     DESC
    0.75
    Marcel
    0.75
     Wrapped
    0.74
     صدیق
    0.74
     المتع
    0.72
    Backup
    0.72
    Suppl
    0.71
    Activations Density 0.002%

    No Known Activations