© 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. 10-GEMMASCOPE-RES-131K
    7. 85387
    Prev
    Next
    INDEX
    Explanations

    identifying 'loader' in code

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

    The neuron fires on code tokens containing the substring “Loader,” effectively spotting occurrences of “Loader” in class, module, or import names.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    瑧
    -2.84
    there
    -2.80
    يست
    -2.77
    and
    -2.66
    d
    -2.63
     innumerable
    -2.59
     jornadas
    -2.56
    卨
    -2.53
    苹
    -2.53
    著名的
    -2.52
    POSITIVE LOGITS
     
    3.31
    部分が
    2.95
    </sub>
    2.83
     "
    2.64
    杁
    2.44
    を選ぶ
    2.36
    部分は
    2.33
     "'
    2.33
    箪
    2.31
     serão
    2.27
    Activations Density 0.003%

    No Known Activations