© 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. 26756
    Prev
    Next
    INDEX
    Explanations

    code syntax or programming language constructs

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

    The neuron fires on numeric literals—tokens consisting of digits (including decimal numbers).

    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
     this
    -1.84
     soyez
    -1.13
     TÉCN
    -1.10
     ürün
    -1.07
     QUÉ
    -1.05
     that
    -1.05
    ﾙ
    -1.05
    ⸒
    -1.02
     rusak
    -1.02
     TÉCNICA
    -1.01
    POSITIVE LOGITS
    所说的
    1.31
     extremely
    1.17
    悻
    1.16
    说的
    1.15
     exceedingly
    1.13
     همچنین
    1.06
    が
    1.06
    における
    1.05
    那個
    1.05
     исключительно
    1.04
    Activations Density 0.002%

    No Known Activations