© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 65910
    Prev
    Next
    INDEX
    Explanations

    specific sentence starts

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

    The neuron detects assembler or linker directives denoted by a leading “.” (e.g. .string, .byte, .align)

    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
     to
    -1.90
     from
    -1.76
     menangis
    -1.75
     in
    -1.70
     getNome
    -1.69
     tertawa
    -1.66
     mendengar
    -1.65
     tersenyum
    -1.57
     sedih
    -1.57
     menik
    -1.55
    POSITIVE LOGITS
    na
    1.84
    el
    1.69
     speziell
    1.56
    is
    1.52
    ar
    1.51
    ur
    1.49
    He
    1.48
     pratiquement
    1.48
     determinados
    1.48
     determinadas
    1.47
    Activations Density 0.000%

    No Known Activations