© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 26214
    Prev
    Next
    INDEX
    Explanations

    but and however

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

    This neuron detects contrastive discourse markers—especially the words “but” or “however” (and the start of the contrasting clause that follows).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ادم
    -1.03
    يف
    -1.03
    vieren
    -0.99
    やっぱり
    -0.99
    اير
    -0.94
    برت
    -0.92
    ཫ
    -0.90
    يت
    -0.89
    льную
    -0.88
    ження
    -0.88
    POSITIVE LOGITS
     can
    1.49
     you
    1.17
    这种情况
    1.09
     otherwise
    1.07
     will
    1.05
     able
    0.98
     tersebut
    0.97
     afford
    0.96
     przecież
    0.96
     free
    0.96
    Activations Density 0.025%

    No Known Activations