© 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. 82822
    Prev
    Next
    INDEX
    Explanations

    exceptional qualities

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

    The neuron chiefly fires on marketing‐style promotional feature phrases—i.e. sales copy praising product qualities (e.g. “Quick, reliable product marking,” “the perfect solution,” “offer safety and reliability,” etc.).

    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
    Folge
    -1.13
     некоторые
    -1.02
    Nachteile
    -1.01
     hermosas
    -0.99
    Erfahrung
    -0.98
     некоторых
    -0.98
    Schicken
    -0.98
     пыта
    -0.95
     sommige
    -0.95
     abbastanza
    -0.94
    POSITIVE LOGITS
     even
    1.68
     particularly
    1.29
     thanks
    1.29
     Correspond
    1.05
    even
    1.04
     оптима
    1.03
     EVEN
    1.02
     convinces
    1.01
     optimal
    1.01
    即使
    1.00
    Activations Density 0.046%

    No Known Activations