© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 14018
    Prev
    Next
    INDEX
    Explanations

    xhtml dtds

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

    This neuron fires on numeric tokens—especially version numbers and other digits (e.g. “4”, “1.1”, “4.0.6”) in technical or code‐like contexts.

    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
    さぁ
    -0.96
    あぁ
    -0.95
     embarazada
    -0.93
    ことにしました
    -0.93
    でご
    -0.93
     råd
    -0.93
    さてさて
    -0.92
     bewust
    -0.90
    こっち
    -0.90
    上班
    -0.89
    POSITIVE LOGITS
    楯
    0.87
    り返
    0.86
    齟
    0.86
    resizingMask
    0.85
     dej
    0.85
    風景
    0.84
     finally
    0.84
     atlet
    0.83
    粋
    0.83
     otto
    0.83
    Activations Density 0.041%

    No Known Activations