© 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. 65431
    Prev
    Next
    INDEX
    Explanations

    public interface

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

    The neuron is sensitive to structural document tokens—things like punctuation marks, formatting or markup symbols (e.g. “<”, “>”, “=”), and other metadata cues that signal section breaks or code/HTML snippets.

    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
     Beek
    -0.91
     cualquier
    -0.90
    📬
    -0.88
     разных
    -0.87
     ſer
    -0.84
     marchés
    -0.83
    adla
    -0.82
     ControllerBase
    -0.81
     gekenn
    -0.81
    ERING
    -0.80
    POSITIVE LOGITS
     other
    0.95
     Ebenso
    0.88
     additional
    0.76
    Junk
    0.75
     vice
    0.75
     even
    0.75
     гале
    0.73
     continuing
    0.73
    ///**
    0.72
     at
    0.71
    Activations Density 0.108%

    No Known Activations