© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5279
    Prev
    Next
    INDEX
    Explanations

    there was/be

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

    The neuron detects prominent interjectional error or notice phrases (like “Oops,” “Sorry,” or “There was”) that signal an apology or error message.

    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
    🏆
    0.62
    ”]
    0.56
    ↵↵↵↵↵
    0.55
    Rosen
    0.54
    Cases
    0.54
    obtain
    0.54
     boston
    0.53
    ↵↵↵↵
    0.51
     ছাড়
    0.51
     nhé
    0.51
    POSITIVE LOGITS
     言っ
    0.59
     supplic
    0.57
    सीम
    0.56
    ปฏิ
    0.55
     行っ
    0.55
     नॉलेज
    0.54
     Moonlight
    0.54
    ualaikum
    0.54
    mvc
    0.53
    粘贴
    0.53
    Activations Density 0.000%

    No Known Activations