© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 75522
    Prev
    Next
    INDEX
    Explanations

    secret

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

    The neuron detects multiword titles or proper‐noun phrases beginning with “Secret,” such as “Secret Love in Peach Blossom Land,” “Divine Secrets of the Ya-Ya Sisterhood,” or “Secret Worlds with Michael Arbuthnot.”

    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
    êtements
    -0.90
     पी
    -0.86
     keadaan
    -0.80
     гель
    -0.78
     edil
    -0.76
     соло
    -0.75
    Est
    -0.74
     patron
    -0.73
     Роди
    -0.72
    IND
    -0.72
    POSITIVE LOGITS
     Secret
    0.93
     Geheimnis
    0.78
     verdadeiro
    0.77
     SECRET
    0.75
     hvert
    0.75
     secret
    0.74
    RunWith
    0.73
    OBJET
    0.72
    Secret
    0.71
    謗
    0.71
    Activations Density 0.015%

    No Known Activations