© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10052
    Prev
    Next
    INDEX
    Explanations

    Bel, Belgium, beluga, bell

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

    The neuron strongly activates on tokens containing the string “Bel,” especially at the start of names or place names.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    the token "Bel" when it appears as part of proper nouns (names of people, places, or organizations).

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    proper nouns that start with the prefix “Bel-” (including place names and surnames).

    oai_token-act-pair · gpt-5Triggered 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
     Cyst
    0.57
     Penit
    0.55
    久
    0.51
    рила
    0.50
    λλα
    0.49
     Kurz
    0.49
     Prost
    0.47
    сет
    0.47
    לת
    0.46
     QUILL
    0.46
    POSITIVE LOGITS
    vedere
    0.83
    bel
    0.72
     Bel
    0.70
    Bel
    0.70
    tran
    0.69
     Бел
    0.67
     Belgian
    0.66
     bel
    0.66
     벨
    0.66
    avista
    0.66
    Activations Density 0.093%

    No Known Activations