© 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. 34-GEMMASCOPE-RES-131K
    7. 86800
    Prev
    Next
    INDEX
    Explanations

    course type, metal company, family

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

    The neuron selectively activates on long, domain-specific single tokens—e.g. class or type identifiers in code (like “TrainingCourse”) or specialized scientific terms—rather than on ordinary words.

    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
     quite
    -1.01
     thinks
    -0.95
     or
    -0.95
     believes
    -0.93
     perhaps
    -0.91
     when
    -0.91
     Quite
    -0.87
     considers
    -0.85
     well
    -0.84
     Perhaps
    -0.84
    POSITIVE LOGITS
    significantly
    0.94
     alho
    0.91
     těchto
    0.91
    摃
    0.89
    
    0.88
    这两
    0.86
    ఝ
    0.86
     Muchas
    0.86
    represents
    0.85
     решений
    0.84
    Activations Density 0.001%

    No Known Activations