© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 15618
    Prev
    Next
    INDEX
    Explanations

    Error messages and types

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

    The neuron activates on tokens typical of program error messages, warnings, and log entries—i.e. pieces of text reporting failures, exceptions, or diagnostic information.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_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
    0.72
    }
    0.69
    зар
    0.67
    ก็จะ
    0.66
    }}=
    0.63
    하기도
    0.63
    AndTime
    0.63
    จะมี
    0.62
     mencionados
    0.62
    얘
    0.61
    POSITIVE LOGITS
     ولم
    0.80
     不能
    0.80
     ڈیزائن
    0.73
     ਨਹੀਂ
    0.72
    <unused18>
    0.71
     நிறுவ
    0.71
     Shrimp
    0.71
    <unused73>
    0.69
    ಚಿತ
    0.69
    ेटिक
    0.69
    Activations Density 0.002%

    No Known Activations