© 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. 72627
    Prev
    Next
    INDEX
    Explanations

    code snippets

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

    This neuron selectively activates on programming‐style tokens (e.g. preprocessor directives like `#ifndef`/`#define`, HTML tags such as `<head>`, Java identifiers like `System`, etc.), i.e. bits of source code or markup.

    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
     zoude
    -1.34
     zwi
    -1.23
     damen
    -1.22
     kompon
    -1.20
     malen
    -1.18
     dezelve
    -1.18
     kram
    -1.16
     staffel
    -1.15
     pérd
    -1.11
     kapture
    -1.11
    POSITIVE LOGITS
     of
    1.00
     that
    0.99
     in
    0.98
    水が
    0.87
     new
    0.83
    ing
    0.81
     under
    0.80
    山頂
    0.80
    紅葉
    0.80
    avan
    0.79
    Activations Density 0.004%

    No Known Activations