© 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. 34-GEMMASCOPE-RES-131K
    7. 62820
    Prev
    Next
    INDEX
    Explanations

    storage locations or file paths

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

    The neuron predominantly activates on code-style tokens and syntax (e.g. SQL keywords, punctuation, file paths, numbers)—in other words, it detects code snippets or code-like formatting.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     CAP
    -0.84
     Smiley
    -0.77
     RUM
    -0.75
    Razer
    -0.71
    isSet
    -0.70
    Steady
    -0.70
     reven
    -0.68
    SOLO
    -0.68
    chevêque
    -0.67
    Memoria
    -0.67
    POSITIVE LOGITS
     bucket
    1.67
    Bucket
    1.38
     Bucket
    1.35
    bucket
    1.34
     buckets
    1.27
    BUCKET
    1.18
     BUCK
    1.12
     gs
    1.09
    桶
    1.03
    Uri
    0.98
    Activations Density 0.040%

    No Known Activations