© 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. 22-GEMMASCOPE-RES-131K
    7. 128789
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on verbs denoting concrete physical actions or handling steps (e.g. pulling, cutting, releasing, unrolling).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    releasing or removing things

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     csomag
    -0.98
    درمان
    -0.97
     ezen
    -0.94
    Tierra
    -0.94
     saludable
    -0.92
    包裹
    -0.90
    uding
    -0.88
    parametrize
    -0.88
     tuturor
    -0.87
    浑身
    -0.86
    POSITIVE LOGITS
     when
    1.95
     during
    1.26
     whenever
    1.13
     когда
    1.12
    When
    1.11
    when
    1.11
     quando
    1.11
     or
    1.08
     use
    1.05
     to
    1.02
    Activations Density 0.118%

    No Known Activations