© 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. 65066
    Prev
    Next
    INDEX
    Explanations

    programming context, specific values

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

    The neuron fires on tokens that aren’t ordinary lowercase words—specifically standalone numbers, uppercase abbreviations (or acronyms), and punctuation/symbol tokens.

    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
     appelés
    -1.16
     gånger
    -1.03
    ならば
    -1.03
     aprovecha
    -1.02
    Actually
    -1.02
    さで
    -1.02
     auxquels
    -1.00
     cuja
    -1.00
     blz
    -1.00
     month
    -0.98
    POSITIVE LOGITS
     possibly
    1.19
     potentially
    1.16
     something
    1.16
     sesuatu
    1.09
    ModelAndView
    1.08
     reduces
    1.05
     edging
    1.04
     conceivably
    1.04
     eliminates
    1.04
     suggests
    1.04
    Activations Density 0.076%

    No Known Activations