© 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. 128736
    Prev
    Next
    INDEX
    Explanations

    This neuron responds to angle-bracket tokens associated with C++ template instantiations (e.g., the `<` and `>` in `shared_ptr<WBAES>`).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    Code snippets, technical terms, names

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     destes
    -1.13
     demuestra
    -1.12
     inspiração
    -1.07
     mehreren
    -1.05
     competição
    -1.05
     mudou
    -1.01
    轻声
    -1.00
     jesteśmy
    -1.00
     jednocześnie
    -1.00
     reduziert
    -0.99
    POSITIVE LOGITS
     However
    1.26
    بعد
    1.17
     vod
    1.13
    闈
    1.13
    بله
    1.09
     浴室
    1.09
     suna
    1.09
     μπορεί
    1.08
     bomba
    1.07
     samo
    1.06
    Activations Density 0.006%

    No Known Activations