© 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. 33377
    Prev
    Next
    INDEX
    Explanations

    aggressive, brutal, harsh

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

    The neuron detects strongly negative, high-intensity adjectives that describe violence or extreme harshness.

    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
    焄
    -2.30
    錚
    -2.23
    ?
    -2.22
    剮
    -2.02
    你会
    -1.96
    )();
    -1.91
    '}}>
    -1.88
    灃
    -1.85
    廕
    -1.82
    鉚
    -1.81
    POSITIVE LOGITS
    6
    2.70
    7
    2.61
    0
    2.55
    8
    2.53
    1
    2.19
    齶
    2.11
    But
    2.06
    3
    2.02
    ness
    1.97
    4
    1.95
    Activations Density 0.019%

    No Known Activations