© 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. 117454
    Prev
    Next
    INDEX
    Explanations

    coordinated serial attacks

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

    The neuron fires whenever it sees named entities—especially place names, proper nouns, and the punctuation (like colons) that typically introduce them in headlines or news‐style text.

    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
     Qatar
    -0.86
    Qatar
    -0.82
     エア
    -0.77
    合わせ
    -0.70
    ambar
    -0.69
     brasileiro
    -0.69
    appel
    -0.68
    ennie
    -0.68
     crops
    -0.67
    繼
    -0.67
    POSITIVE LOGITS
     coordinated
    1.45
     serial
    1.10
     attacks
    1.07
     COORDIN
    1.05
     Coordin
    1.03
     multiple
    0.93
    coord
    0.93
     coordination
    0.91
     coordon
    0.91
     coordin
    0.88
    Activations Density 0.029%

    No Known Activations