© 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. 7554
    Prev
    Next
    INDEX
    Explanations

    when discussion/dealing

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

    This neuron detects the discourse phrase “when it comes to” (and its variants like “it comes to”) that signals introducing a topic.

    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
     прошед
    0.54
     fil
    0.54
    sel
    0.53
    λία
    0.52
     пройдет
    0.52
     sel
    0.52
     Counsel
    0.52
    д
    0.51
    לי
    0.50
     Kennedy
    0.50
    POSITIVE LOGITS
     chodzi
    1.08
     comes
    1.07
    涉及到
    1.06
     matter
    0.99
     речь
    0.97
     cuestión
    0.97
    涉及
    0.91
     يتعلق
    0.91
     касается
    0.89
     dealing
    0.88
    Activations Density 0.173%

    No Known Activations