© 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. 47708
    Prev
    Next
    INDEX
    Explanations

    should / should not

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

    This neuron is detecting runs of leading whitespace (indentation) at the start of code lines.

    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
     Best
    -0.84
     aveva
    -0.78
    要不是
    -0.77
     già
    -0.75
     best
    -0.75
     besoin
    -0.73
     Knew
    -0.73
    🥖
    -0.71
     함께
    -0.71
    best
    -0.70
    POSITIVE LOGITS
     should
    1.82
     shall
    1.39
    should
    1.38
     shouldn
    1.30
    應該
    1.13
     должна
    1.11
     должен
    1.06
     должно
    1.05
     powinno
    1.02
     shouldBe
    1.02
    Activations Density 0.007%

    No Known Activations