© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 13115
    Prev
    Next
    INDEX
    Explanations

    adjective noun combinations

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

    The neuron strongly activates on domain-specific or technical terms (e.g. precise jargon like “getParameterValues,” “architecture,” “multi,” “negative,” “impartial,” etc.).

    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
     ciga
    -1.00
     perfeito
    -1.00
     bumi
    -0.98
     cruze
    -0.94
     pite
    -0.93
     bailey
    -0.93
    rét
    -0.90
     kev
    -0.88
     gonz
    -0.87
    Certain
    -0.85
    POSITIVE LOGITS
     détru
    0.95
    itaries
    0.93
     damaligen
    0.90
    bih
    0.89
     By
    0.87
    ходил
    0.84
     vieron
    0.84
     condensate
    0.83
     expertos
    0.83
     (!)
    0.82
    Activations Density 0.173%

    No Known Activations