© 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. 67592
    Prev
    Next
    INDEX
    Explanations

    groups and combinations

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

    The neuron selectively activates on long, content‐heavy technical nouns and jargon (e.g. “patterns,” “groups,” “dependence,” “coefficients”) common in scientific abstracts.

    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
    dampak
    -0.96
    middels
    -0.88
    品的
    -0.85
    elstein
    -0.82
     JUL
    -0.81
    EACH
    -0.80
     the
    -0.79
    acchar
    -0.77
    跟着
    -0.77
     Þ
    -0.77
    POSITIVE LOGITS
     group
    1.18
     combinations
    1.18
     groups
    1.16
     coworkers
    1.05
     composed
    1.05
    ()>
    1.03
    dessus
    1.01
     arreglos
    1.00
     teammates
    1.00
    %。
    1.00
    Activations Density 0.112%

    No Known Activations