© 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. 128601
    Prev
    Next
    INDEX
    Explanations

    The neuron responds to standalone numeric tokens (integer and floating‐point literals) in code and diff metadata.

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

    #ifndef or diff --git

    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
    電車
    -0.97
     while
    -0.92
    стик
    -0.88
     only
    -0.84
     FixedUpdate
    -0.83
     reckless
    -0.82
    だからこそ
    -0.82
    Технические
    -0.82
     incessant
    -0.81
    ="/">
    -0.78
    POSITIVE LOGITS
     such
    1.05
    になるので
    0.96
     skriv
    0.93
     такого
    0.92
     croce
    0.89
     aient
    0.89
     bepa
    0.89
     arquivo
    0.88
     esimer
    0.88
     australiano
    0.88
    Activations Density 0.047%

    No Known Activations