© 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. 31-GEMMASCOPE-2-RES-16K
    7. 15903
    Prev
    Next
    INDEX
    Explanations

    cannot / negation

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

    The neuron fires strongly on each snippet’s single most salient content word—effectively picking out the one key or emphasis-bearing token in a span.

    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
    有机会
    0.90
    有可能
    0.72
     mogelijk
    0.68
    可能性
    0.65
     способности
    0.65
    的可能性
    0.64
     mulig
    0.64
     possibilité
    0.63
     ability
    0.63
     riesce
    0.63
    POSITIVE LOGITS
    cannot
    2.77
     cannot
    2.72
     Cannot
    2.56
    Cannot
    2.50
     нельзя
    2.45
    不能
    2.25
     不能
    2.03
    不可以
    1.96
    不可
    1.90
    无法
    1.90
    Activations Density 0.404%

    No Known Activations