© 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. 10-GEMMASCOPE-RES-131K
    7. 69189
    Prev
    Next
    INDEX
    Explanations

    despite progress or success

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

    This neuron strongly activates on very common, high-frequency words (mostly function words and ubiquitous academic terms) rather than content-specific terms.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     when
    -2.47
     is
    -2.16
     instead
    -2.00
     not
    -2.00
    </i>
    -1.80
     where
    -1.69
     before
    -1.66
     rather
    -1.66
     so
    -1.55
     what
    -1.55
    POSITIVE LOGITS
     doute
    1.55
     tromper
    1.54
     rafra
    1.52
     enthou
    1.48
     içi
    1.48
     muñ
    1.47
     komplet
    1.45
     sabia
    1.45
     pedimos
    1.45
     sahibi
    1.44
    Activations Density 0.052%

    No Known Activations