© 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. 6122
    Prev
    Next
    INDEX
    Explanations

    but here

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

    The neuron selectively fires on discourse-connective tokens (e.g. conjunctions and transitional words like “and,” “but,” “if,” “although,” “firstly,” etc.).

    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
    왜
    0.59
    Quién
    0.56
     Why
    0.55
    誰も
    0.54
     Who
    0.51
     why
    0.50
    πο
    0.50
    Why
    0.49
    Who
    0.49
     왜
    0.49
    POSITIVE LOGITS
     everytime
    0.85
     whilst
    0.81
     although
    0.78
     aunque
    0.76
     meskipun
    0.75
     essendo
    0.74
     enquanto
    0.73
     while
    0.72
    雖然
    0.70
     walaupun
    0.70
    Activations Density 0.158%

    No Known Activations