© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 85717
    Prev
    Next
    INDEX
    Explanations

    man who or who

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

    The neuron responds to title‐case or headline-style capitalized words (e.g. “Man,” “Who,” “Killed”) indicating tokens in a news headline or article title.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     innych
    -1.93
     colegios
    -1.92
     dinosaurio
    -1.82
     envíos
    -1.80
     neumáticos
    -1.78
     muelle
    -1.77
     hospitales
    -1.75
    idee
    -1.73
     carteles
    -1.72
     conductores
    -1.70
    POSITIVE LOGITS
     This
    1.73
     In
    1.67
    /
    1.64
    哎呀
    1.57
     их
    1.56
     Many
    1.55
     atau
    1.52
     While
    1.51
    .
    1.50
     their
    1.48
    Activations Density 0.052%

    No Known Activations