© 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. 59281
    Prev
    Next
    INDEX
    Explanations

    variable names and punctuation

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

    The neuron fires on programming‐language identifiers for data objects—e.g. variable names and objects whose methods or properties (like `.extend`, `.fetchall`, error lists, data frames, etc.) are being accessed.

    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
     or
    -2.63
     when
    -2.47
     without
    -2.45
     does
    -2.38
     if
    -2.38
     first
    -2.36
     only
    -2.30
     this
    -2.25
     gets
    -2.11
     immediately
    -2.11
    POSITIVE LOGITS
    ཿ
    1.80
    unfortunately
    1.77
    傒
    1.77
    ≃
    1.76
     severa
    1.75
    idk
    1.73
    鲼
    1.73
    
    1.73
     suppos
    1.73
     acheter
    1.73
    Activations Density 0.018%

    No Known Activations