© 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. 7224
    Prev
    Next
    INDEX
    Explanations

    negations and contractions (isn't, didn't)

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

    The neuron lights up on low‐frequency, content-heavy tokens—essentially rarer or more domain-specific words rather than common function words.

    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
     மற்றும்
    0.81
    +
    0.75
    にて
    0.71
     ২২
    0.71
     सहित
    0.70
     junto
    0.70
     adlı
    0.70
     ৩০
    0.68
    /
    0.67
    および
    0.66
    POSITIVE LOGITS
     Shouldn
    1.03
     Nobody
    0.98
     Wouldn
    0.97
     오히려
    0.97
     Policing
    0.96
    即使
    0.96
     Даже
    0.95
    Doesn
    0.95
     niemand
    0.93
     Ведь
    0.92
    Activations Density 0.187%

    No Known Activations