© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 86425
    Prev
    Next
    INDEX
    Explanations

    programming variables and keywords

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

    This neuron mainly detects the punctuation of function/ method argument lists (especially opening parentheses and commas).

    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
     soooo
    -1.99
    绺
    -1.91
     spearheaded
    -1.91
     myriad
    -1.90
    辔
    -1.88
     diverses
    -1.88
     wacky
    -1.85
     parfois
    -1.84
    EVERY
    -1.83
    seems
    -1.83
    POSITIVE LOGITS
     first
    2.20
     enormous
    2.19
     increase
    2.16
     three
    2.13
    With
    2.09
     information
    2.05
    Before
    2.03
     Knowing
    2.03
     methods
    2.03
     provide
    2.03
    Activations Density 0.032%

    No Known Activations