© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 50295
    Prev
    Next
    INDEX
    Explanations

    javadoc build version

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

    The neuron is picking out code‐metadata and header/declaration tokens—things like assembly-attribute lines, module/license headers, version/build numbers, angle-bracketed tags, and similar file-header syntax.

    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
     Collected
    -0.72
    œuf
    -0.70
     instituição
    -0.68
    螂
    -0.66
    Tanks
    -0.66
     based
    -0.66
    하십시오
    -0.66
    彙
    -0.66
    有问题
    -0.65
     studios
    -0.65
    POSITIVE LOGITS
    englisch
    0.75
    Sight
    0.74
    gebnis
    0.74
    ↑
    0.66
     ۵
    0.66
    packageName
    0.66
    升
    0.66
     wakati
    0.66
     گروه
    0.65
    sight
    0.65
    Activations Density 0.073%

    No Known Activations