© 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. 65675
    Prev
    Next
    INDEX
    Explanations

    code syntax and constructs

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

    The neuron is primarily triggered by numeric literals and version‐style tokens (e.g. numbers, version strings, and related punctuation) in code.

    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
     BELIEVE
    -0.94
     tslint
    -0.91
     Ficha
    -0.89
    iens
    -0.88
     భ
    -0.84
    deren
    -0.84
    宍
    -0.83
    艿
    -0.82
    ]+=
    -0.81
     chocol
    -0.80
    POSITIVE LOGITS
     if
    1.25
     or
    1.02
     arba
    0.94
     override
    0.91
     that
    0.89
     jeśli
    0.88
     If
    0.88
     trinken
    0.87
     when
    0.84
     saját
    0.83
    Activations Density 0.017%

    No Known Activations