© 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. 22-GEMMASCOPE-RES-131K
    4. 68336
    Prev
    Next
    INDEX
    Explanations

    first person statements

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

    The neuron is strongly triggered by conversational framing words—especially first-person/second-person pronouns and polite acknowledgments (e.g. I, you, thank, glad, congrats)—that introduce or respond in forum-style posts.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     acides
    -1.28
    -¡
    -1.09
     colonias
    -1.07
    Jawab
    -1.03
    fieldId
    -1.00
    herstellung
    -0.98
    ↵↵↵
    -0.98
     tages
    -0.97
    」,
    -0.96
    jTable
    -0.95
    POSITIVE LOGITS
     you
    1.88
     your
    1.72
     that
    1.10
    と思いますが
    1.03
    Referenties
    0.99
     уже
    0.98
     what
    0.98
     would
    0.98
     can
    0.96
     скорее
    0.95
    Activations Density 0.015%

    No Known Activations