© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72840
    Prev
    Next
    INDEX
    Explanations

    tracking, trials, statistics

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

    The neuron preferentially fires on low‐frequency or uncommon tokens (i.e. rare vocabulary items).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Japon
    -1.01
     berbeda
    -1.00
     parche
    -1.00
     tille
    -0.99
     itali
    -0.97
     vinil
    -0.97
     laiton
    -0.96
     laine
    -0.94
    Kategori
    -0.94
     essentielle
    -0.94
    POSITIVE LOGITS
     cortesía
    0.85
     —
    0.82
    熟悉的
    0.82
    a
    0.79
     \{
    0.77
     fuf
    0.76
    p
    0.76
    _]
    0.76
    X
    0.76
    Q
    0.76
    Activations Density 0.077%

    No Known Activations