© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 6530
    Prev
    Next
    INDEX
    Explanations

    drums and music introduction

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

    The neuron fires on multi‐word title‐case or headline‐style phrases (e.g. “Quality First…,” “Retrieved from,” “ELF Object Files,” 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
    Pss
    0.72
     proceeding
    0.69
     nearby
    0.65
    ⁈
    0.65
    堂
    0.64
    图中
    0.64
     forse
    0.63
    onk
    0.63
    ℠
    0.63
    哿
    0.62
    POSITIVE LOGITS
     após
    0.95
     alegre
    0.94
     começa
    0.88
     elétrica
    0.87
     inicia
    0.87
     podendo
    0.86
     начинает
    0.86
     FAILURE
    0.86
     பெரும்பாலான
    0.84
    い
    0.83
    Activations Density 0.001%

    No Known Activations