© 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. 34-GEMMASCOPE-RES-131K
    7. 99263
    Prev
    Next
    INDEX
    Explanations

    subject of debate or study

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

    This neuron fires on occurrences of “subject” (and similar topic‐of‐discussion words) as in “the subject of…” to mark mentions of what a text is about.

    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
    脸上的
    -0.91
    BUT
    -0.90
    idée
    -0.88
    TagMode
    -0.88
     Rook
    -0.84
     idea
    -0.84
    HasForeignKey
    -0.83
    wea
    -0.82
    CREDIT
    -0.82
    样本
    -0.81
    POSITIVE LOGITS
     attention
    1.63
     much
    1.52
     of
    1.39
     intense
    1.39
     interest
    1.33
     discussion
    1.32
     considerable
    1.30
     study
    1.29
     focus
    1.28
     concern
    1.27
    Activations Density 0.037%

    No Known Activations