© 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. 25864
    Prev
    Next
    INDEX
    Explanations

    private and readonly declarations

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

    The neuron fires on programming‐language syntax tokens (e.g. punctuation like “(”, “,”, “:”, keywords in code contexts), i.e. it detects code-style structural elements rather than natural-language words.

    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
    Specifik
    -1.23
     comté
    -1.22
     bibit
    -1.18
     consag
    -1.15
    Hrs
    -1.14
     atleta
    -1.13
    Groet
    -1.12
     prêtre
    -1.09
     hatta
    -1.09
    шкина
    -1.08
    POSITIVE LOGITS
     system
    1.08
     набор
    1.03
     resources
    1.02
     access
    1.00
    farb
    0.99
     performing
    0.97
    }}_{\
    0.97
     performed
    0.97
     service
    0.96
     and
    0.96
    Activations Density 0.016%

    No Known Activations