© 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. 22-GEMMASCOPE-RES-131K
    7. 28069
    Prev
    Next
    INDEX
    Explanations

    invalid or not found messages

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

    The neuron fires on words and tokens that appear in error or failure messages (e.g. “is an invalid…”, “Sorry, this video is…”, “version is not set”), i.e. it detects error-reporting language.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    μών
    -1.17
    therners
    -0.98
    guh
    -0.98
    utiliser
    -0.98
    Ejemplos
    -0.97
    nasi
    -0.95
    endiendo
    -0.94
    
    -0.94
    膈
    -0.93
     rød
    -0.93
    POSITIVE LOGITS
     or
    1.83
     not
    1.77
     either
    1.62
     либо
    1.27
     neither
    1.16
     perhaps
    1.16
     insufficient
    1.16
    或是
    1.15
     didn
    1.13
     wasn
    1.12
    Activations Density 0.005%

    No Known Activations