© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 2964
    Prev
    Next
    INDEX
    Explanations

    The neuron detects first‐person endorsement phrases—especially “I recommend,” “highly recommend,” “I say,” etc., that express personal recommendation.

    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_24_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
    왔
    0.79
    Fecha
    0.71
    GAN
    0.66
    مبر
    0.66
    Misc
    0.66
    WebServer
    0.66
    うまく
    0.65
    뭐
    0.64
    ElapsedTime
    0.64
    BA
    0.62
    POSITIVE LOGITS
     wholeheartedly
    0.98
     strongly
    0.91
     Ihnen
    0.89
     sangat
    0.88
     heartily
    0.87
     вас
    0.86
     fortement
    0.86
     કોઈપણ
    0.85
     σας
    0.84
     ustedes
    0.84
    Activations Density 0.088%

    No Known Activations