© 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. 13209
    Prev
    Next
    INDEX
    Explanations

    The neuron is specialized to pick up the existential “there was a/an” construction in text.

    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
    y
    0.61
     training
    0.57
     vanish
    0.55
    x
    0.54
    হের
    0.54
     jobs
    0.53
     tus
    0.52
     pods
    0.52
     x
    0.52
     there
    0.51
    POSITIVE LOGITS
    ску
    0.67
     Україні
    0.61
    コネ
    0.60
     लागले
    0.59
     Рей
    0.58
    接続
    0.58
    evším
    0.57
     माध्यमातून
    0.56
    ărul
    0.56
    印刷
    0.55
    Activations Density 0.000%

    No Known Activations