© 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. 34-GEMMASCOPE-RES-131K
    7. 2568
    Prev
    Next
    INDEX
    Explanations

    Java class debugging

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

    The neuron strongly activates on programming-style tokens—identifiers, class/type names, package paths, and other code-like jargon—i.e. it is looking for snippets of source code or code-related terminology.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    existsSync
    -0.98
    makeText
    -0.96
    ErrorException
    -0.94
    MethodManager
    -0.91
    恻
    -0.91
    setArguments
    -0.91
    getFullYear
    -0.90
    ERICAL
    -0.90
    angka
    -0.90
    FirstResponder
    -0.90
    POSITIVE LOGITS
     java
    1.55
     javax
    1.39
     {@
    1.29
     clazz
    1.18
    java
    1.15
     Java
    1.13
     JDK
    1.11
     JVM
    1.08
    jspx
    1.03
    javax
    1.02
    Activations Density 0.590%

    No Known Activations