© 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. 3421
    Prev
    Next
    INDEX
    Explanations

    a capability

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

    The neuron strongly responds to low‐frequency, content‐bearing words (i.e. uncommon or high–information‐content tokens).

    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
    aquest
    0.62
    tional
    0.61
    0.60
    pearl
    0.59
    s
    0.58
    ebook
    0.58
    _
    0.57
     м
    0.57
    ife
    0.57
    athon
    0.57
    POSITIVE LOGITS
     bedeutet
    1.05
     innebär
    1.02
     implies
    1.00
     equates
    0.96
     necessitates
    0.95
     precludes
    0.94
     činjen
    0.93
     betekent
    0.93
     contradicts
    0.93
     Tatsache
    0.89
    Activations Density 0.506%

    No Known Activations