© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 39431
    Prev
    Next
    INDEX
    Explanations

    physical combat and abilities

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

    The neuron fires on words naming combat attack types and distances—especially “melee,” “ranged,” “missiles,” and “range.”

    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
     in
    -2.63
     on
    -1.95
     that
    -1.48
     where
    -1.34
     new
    -1.32
     about
    -1.32
     did
    -1.30
     asked
    -1.29
     while
    -1.27
     went
    -1.23
    POSITIVE LOGITS
    苤
    1.34
    zula
    1.30
    reis
    1.27
     constater
    1.23
     devront
    1.22
     charmant
    1.22
    azia
    1.21
     soldat
    1.20
    sieren
    1.19
    κος
    1.19
    Activations Density 0.033%

    No Known Activations