© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 66300
    Prev
    Next
    INDEX
    Explanations

    problems and solutions

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

    The neuron is selectively activating on open‐class content words (nouns, verbs, adjectives, etc.) rather than on function 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
     inexorable
    -1.06
    tauscht
    -0.97
     inclusief
    -0.95
    为此
    -0.94
     therefor
    -0.92
     reconno
    -0.91
     Capsules
    -0.91
    数値
    -0.90
     sanguine
    -0.88
     fhe
    -0.87
    POSITIVE LOGITS
     czyli
    1.16
     what
    1.05
    setLayoutManager
    1.00
    uluş
    0.99
     menthe
    0.98
    érc
    0.98
     ),
    0.97
     the
    0.97
    matician
    0.96
    وفة
    0.96
    Activations Density 0.049%

    No Known Activations