© 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. 36160
    Prev
    Next
    INDEX
    Explanations

    actions followed by completion

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

    The neuron strongly activates on very common English function words—short, high‐frequency articles, prepositions, and auxiliaries such as “the,” “of,” “to,” and “is.”

    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
    före
    -1.76
     parceiros
    -1.46
    områ
    -1.45
    vård
    -1.40
    när
    -1.38
    都需要
    -1.34
    fören
    -1.30
    诡异
    -1.29
    område
    -1.27
     förra
    -1.26
    POSITIVE LOGITS
     evidence
    1.46
     options
    1.18
    ització
    1.17
    無し
    1.16
    paket
    1.16
     protecting
    1.16
     nutzt
    1.15
    ateral
    1.13
     protects
    1.13
    uating
    1.11
    Activations Density 0.010%

    No Known Activations