© 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. 22-GEMMASCOPE-RES-131K
    7. 39823
    Prev
    Next
    INDEX
    Explanations

    years, versions, dates

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

    The neuron fires almost exclusively on non-natural-language tokens—things like version numbers, numeric values, HTML/XML tags, section headers, and other “structural” or metadata tokens rather than 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_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
     GENT
    -0.86
    ceous
    -0.84
     Grenville
    -0.82
     Clapton
    -0.82
    ATED
    -0.82
    published
    -0.81
     Résultat
    -0.81
    鱒
    -0.80
     Trouver
    -0.80
     exor
    -0.79
    POSITIVE LOGITS
     maraming
    0.76
    無
    0.76
     IERC
    0.75
     卧室
    0.73
    weli
    0.72
    无
    0.72
    érios
    0.70
    timb
    0.70
    Adapters
    0.70
     无
    0.69
    Activations Density 0.007%

    No Known Activations