© 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. 1874
    Prev
    Next
    INDEX
    Explanations

    actions causing notifications or errors

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

    The neuron fires on tokens related to error and exception handling in code snippets (e.g. “raise,” “exception,” “error,” “abort,” “returned,” “print a message”).

    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
    的一些
    -0.92
    傢
    -0.90
    鐐
    -0.88
     pratiquement
    -0.87
     некоторые
    -0.86
    unexpected
    -0.85
     니
    -0.84
     wretched
    -0.82
    archiwizowane
    -0.82
    Gestion
    -0.81
    POSITIVE LOGITS
     an
    1.80
     соответству
    1.45
     a
    1.39
     appropriate
    1.16
     notification
    1.16
     accordingly
    1.13
     message
    1.08
     entsprechenden
    1.05
     corresponding
    1.01
     respective
    1.00
    Activations Density 0.120%

    No Known Activations