© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 71318
    Prev
    Next
    INDEX
    Explanations

    True, Q:, resource", I have

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

    This neuron fires on programming-language literal tokens that mark truth or block boundaries—specifically the boolean “True” literal and the closing curly brace “}.”

    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
     all
    -1.33
    леге
    -1.21
    𝒙
    -1.18
     встречи
    -1.17
    めます
    -1.13
     feveral
    -1.09
    ritsar
    -1.09
    出声
    -1.09
     Aralık
    -1.09
    埌
    -1.08
    POSITIVE LOGITS
     fleste
    1.48
    giendo
    1.41
     engraçado
    1.38
    versy
    1.37
    Ethylbenzene
    1.34
     meninos
    1.27
    Cependant
    1.27
     but
    1.25
    ngua
    1.23
    ણી
    1.23
    Activations Density 0.000%

    No Known Activations