© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5619
    Prev
    Next
    INDEX
    Explanations

    greeting words

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

    The neuron responds strongly to salutation and valediction words and phrases (greetings like “Hello,” and parting words like “Goodbye,” “Farewell,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Ove
    0.64
     Monique
    0.63
    ধর
    0.61
     Rape
    0.58
     Petitioner
    0.56
    نر
    0.56
    ]{
    0.55
    ಪಡ
    0.55
     Watching
    0.55
     Sheldon
    0.55
    POSITIVE LOGITS
     goodbye
    1.89
    Goodbye
    1.76
     farewell
    1.70
     Goodbye
    1.59
    👋
    1.48
     desped
    1.42
     Farewell
    1.27
     👋
    1.27
     parting
    1.25
     bye
    1.24
    Activations Density 0.192%

    No Known Activations