© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 47355
    Prev
    Next
    INDEX
    Explanations

    steam related compounds and terms

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

    The neuron is highly sensitive to mentions of “steam” (in any context or casing) and activates whenever the token “steam” appears.

    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
    anel
    -0.77
    NSDictionary
    -0.72
     контроля
    -0.72
     shelving
    -0.71
     controls
    -0.70
    一枚
    -0.69
     waveform
    -0.69
    love
    -0.68
    として
    -0.68
     statistique
    -0.68
    POSITIVE LOGITS
     蒸
    1.14
    roller
    1.09
     punk
    1.08
    ROLLER
    1.05
    Punk
    1.01
    powered
    0.96
    Powered
    0.94
    rierten
    0.91
    ulated
    0.91
     Steam
    0.90
    Activations Density 0.010%

    No Known Activations