© 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. 98578
    Prev
    Next
    INDEX
    Explanations

    quantifying size and measurements

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

    The neuron fires on uncommon, document-specific tokens—especially longer or technical words (e.g. function names, file paths, specialized jargon) that only appear once or very infrequently.

    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
     for
    -1.80
     at
    -1.56
     on
    -1.41
     as
    -1.39
     eight
    -1.37
     provide
    -1.33
     any
    -1.30
     five
    -1.30
     nine
    -1.29
     after
    -1.27
    POSITIVE LOGITS
    扨
    1.48
    㩒
    1.34
    めちゃくちゃ
    1.34
     égale
    1.31
     évidemment
    1.29
     insanely
    1.28
     arrêter
    1.26
    賸
    1.25
    
    1.25
     enraged
    1.24
    Activations Density 0.047%

    No Known Activations