© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 9279
    Prev
    Next
    INDEX
    Explanations

    raising or throwing errors

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

    The main thing this neuron does is detect mentions of runtime errors and exception‐handling constructs (e.g. “error,” “exception,” “On Error GoTo,” etc.).

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

    programming error handling contexts, particularly when describing exceptions, error conditions, or error-related code flow.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    technical text about software errors and exceptions, especially error-handling terminology in programming code, documentation, or logs.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     yon
    0.59
     tématu
    0.57
    惦
    0.56
     засе
    0.55
    ümüş
    0.55
    黾
    0.54
     進
    0.54
     காங்க
    0.54
     jobNumber
    0.52
     katol
    0.52
    POSITIVE LOGITS
    throw
    2.11
     throw
    2.03
     throwing
    2.02
     throws
    1.94
    Throw
    1.94
     Throw
    1.91
    throwing
    1.81
    throws
    1.78
     thrown
    1.76
    Throws
    1.74
    Activations Density 0.035%

    No Known Activations