© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 8211
    Prev
    Next
    INDEX
    Explanations

    while contrasting with consequences

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

    This neuron is highly sensitive to the subordinating conjunction “while,” especially when it begins a contrasting clause.

    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_24_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
    ни
    1.10
    м
    1.02
     ardından
    0.85
     然而
    0.84
    後來
    0.82
    ினும்
    0.81
     rapidement
    0.80
     NoSuch
    0.79
    随后
    0.79
     Hinweis
    0.77
    POSITIVE LOGITS
    toler
    0.85
    EAST
    0.84
    에요
    0.83
    obed
    0.80
    ade
    0.80
    %)
    0.79
    TRUE
    0.79
    obje
    0.79
    낸
    0.79
    parametrize
    0.78
    Activations Density 0.039%

    No Known Activations