© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 28566
    Prev
    Next
    INDEX
    Explanations

    coach said he

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

    The neuron detects speaker attributions, i.e. patterns where a person’s name (often a coach) is followed by a reporting verb like “said.”

    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
    檔
    -1.02
    釐
    -1.01
    闡
    -0.99
    Discography
    -0.97
    ACCIÓN
    -0.94
    当時
    -0.92
    bourg
    -0.92
     "",
    -0.91
    calciatore
    -0.91
    izace
    -0.90
    POSITIVE LOGITS
     of
    1.10
     waschen
    1.02
    那些
    1.00
     Toutes
    0.97
     mostaza
    0.96
     cuchara
    0.95
     teaching
    0.91
    sividad
    0.91
     
    0.90
     coaching
    0.90
    Activations Density 0.022%

    No Known Activations