© 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. 10-GEMMASCOPE-RES-131K
    7. 59603
    Prev
    Next
    INDEX
    Explanations

    person to person

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

    The neuron fires on occurrences of “person” when it appears as a hyphenated prefix in compound terms (e.g. “person‐oriented,” “person‐based,” “person‐to‐person”).

    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
     insistent
    -2.02
    Ⴀ
    -1.96
    satunya
    -1.90
    炙
    -1.88
     varandra
    -1.84
     acceso
    -1.83
    τυχ
    -1.80
     principalement
    -1.80
    🫓
    -1.80
     certainement
    -1.77
    POSITIVE LOGITS
    スポンサーリンク
    2.27
     The
    2.19
    というわけで
    1.99
    そういえば
    1.98
    byes
    1.96
    余談
    1.94
    んですが
    1.88
    when
    1.87
    zembro
    1.83
    誰か
    1.82
    Activations Density 0.034%

    No Known Activations