© 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. 34668
    Prev
    Next
    INDEX
    Explanations

    programming type keywords

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

    The neuron strongly activates on numeric literals (digits and numbers) and all‐caps identifier tokens (e.g. version numbers, bit-widths, environment variable names), i.e. code tokens that represent numbers or uppercase constants.

    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
     этих
    -1.13
     cũng
    -0.90
     about
    -0.89
     ซึ่ง
    -0.89
     почти
    -0.84
     noblest
    -0.84
     of
    -0.83
    still
    -0.82
     Miß
    -0.82
     magnificent
    -0.80
    POSITIVE LOGITS
     preferably
    1.27
     facturas
    1.02
     provided
    1.01
    ąp
    0.96
     installiert
    0.93
     cannot
    0.92
    altro
    0.92
     arreg
    0.92
     недорого
    0.91
    才行
    0.89
    Activations Density 0.015%

    No Known Activations