© 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. 5586
    Prev
    Next
    INDEX
    Explanations

    This neuron fires strongly on the final content word just before a segment boundary (i.e. the token immediately preceding an <end> or chunk break).

    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
     เง
    0.68
    Vertical
    0.66
     Periodic
    0.66
    Periodic
    0.66
     съм
    0.65
    Decrease
    0.65
     অ্যাসিড
    0.65
    ᐟ
    0.64
     trustworthiness
    0.64
     Ensuring
    0.64
    POSITIVE LOGITS
     drugiej
    0.82
    ارس
    0.74
     biologist
    0.74
     nagyobb
    0.73
     Jara
    0.73
     drugi
    0.73
     expts
    0.72
     miesią
    0.71
    大型
    0.70
     expt
    0.69
    Activations Density 0.000%

    No Known Activations