© 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. 102892
    Prev
    Next
    INDEX
    Explanations

    ordering words like first and last

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

    The neuron highlights the central action word (the main verb) that names the task the author wants to perform.

    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
    wór
    -1.13
    (!__
    -1.12
    Previous
    -1.05
    如今
    -0.97
     både
    -0.92
    lightbox
    -0.88
    жного
    -0.88
    WAUKEE
    -0.88
     fakta
    -0.88
    autoreleasepool
    -0.87
    POSITIVE LOGITS
     last
    4.94
     first
    4.06
    last
    3.11
     second
    3.03
     LAST
    2.94
     último
    2.75
     послед
    2.67
    Last
    2.66
     terakhir
    2.56
    first
    2.52
    Activations Density 0.063%

    No Known Activations