© 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. 5827
    Prev
    Next
    INDEX
    Explanations

    list names

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

    This neuron reliably spikes on the very first contentful token of each new text segment (i.e. right after a `<start>`), effectively detecting the start of a new document or section.

    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
    на
    1.05
    न
    0.96
    ان
    0.93
    मा
    0.85
    '
    0.85
    ون
    0.81
    Gon
    0.80
    வி
    0.78
    ↵
    0.76
    ম
    0.75
    POSITIVE LOGITS
     exorbit
    0.89
    บคุม
    0.88
    дентификаторы
    0.87
    ENCIA
    0.81
     authorize
    0.80
    ﻯ
    0.80
     enfermos
    0.79
    ంబేద్కర్
    0.78
    手中
    0.77
    孜
    0.77
    Activations Density 0.000%

    No Known Activations