© 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. 13352
    Prev
    Next
    INDEX
    Explanations

    actions performed and their context

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

    The neuron responds to past-tense verbs (especially “-ed” forms and other past participles) that describe actions of creating, building, engineering, or managing something.

    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
     when
    -1.29
     have
    -1.21
     will
    -1.13
     their
    -1.09
     had
    -1.08
     suitable
    -1.07
     has
    -1.02
     because
    -1.00
     began
    -0.98
     achieve
    -0.97
    POSITIVE LOGITS
     pyta
    1.02
     exceedingly
    1.02
    非常的
    0.99
     ΤΟ
    0.96
    does
    0.95
    many
    0.95
    感受到
    0.92
    completely
    0.91
    希望能
    0.91
    quite
    0.90
    Activations Density 0.039%

    No Known Activations