© 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. 34-GEMMASCOPE-RES-131K
    7. 106572
    Prev
    Next
    INDEX
    Explanations

    removing stains and rust

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

    This neuron fires on verbs that describe individual steps in a DIY or cleaning procedure (e.g. brushing, scrubbing, remove, use).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     gluing
    -1.02
     lakier
    -0.96
    шили
    -0.93
    ISTICS
    -0.91
     computadoras
    -0.88
     gemessen
    -0.85
    Chaque
    -0.84
     SOMEONE
    -0.83
    refine
    -0.82
     malpractice
    -0.82
    POSITIVE LOGITS
     tackling
    1.29
     tackled
    1.13
     solutions
    1.05
     homeowners
    1.04
    dymyr
    1.03
     DIY
    1.01
    targeting
    0.96
     you
    0.91
     prevention
    0.91
     tackle
    0.91
    Activations Density 0.010%

    No Known Activations