© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 7500
    Prev
    Next
    INDEX
    Explanations

    positive greetings and acknowledgments

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

    The neuron activates on short positive-feedback words or phrases—e.g. “Good,” “Nice,” “Glad,” “Keep going”—i.e. expressions of praise or encouragement.

    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
    த்துக்கு
    0.67
    regarded
    0.66
    mentare
    0.64
    artition
    0.57
    ंत्रिकी
    0.56
     ሁኔታ
    0.55
     صاح
    0.54
    ᕝ
    0.54
     অনুপ
    0.54
    曾在
    0.53
    POSITIVE LOGITS
     meeting
    0.72
     Meeting
    0.72
     point
    0.64
     중요
    0.64
     Evening
    0.61
     MEETING
    0.59
     morning
    0.59
     Affair
    0.59
     Issue
    0.59
     செயல்
    0.58
    Activations Density 0.057%

    No Known Activations