© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 1767
    Prev
    Next
    INDEX
    Explanations

    prepositions followed by articles or specific nouns

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

    The neuron fires almost exclusively on numeric tokens—especially floating‐point numbers (i.e. decimal numbers).

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

    the word "from" when it appears in contexts describing sources, origins, or starting points of revenue, profit, distance, or other measurable quantities.

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

    prepositional phrases that indicate a source or origin relationship.

    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_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
     kerajaan
    0.62
     pemerintah
    0.57
     transformative
    0.57
     способны
    0.57
     Dhabi
    0.57
     bangsa
    0.57
     mieć
    0.57
     oamenii
    0.56
     cuore
    0.56
     CERN
    0.56
    POSITIVE LOGITS
    each
    0.77
     each
    0.70
    每次
    0.61
    第一次
    0.60
     첫
    0.60
    during
    0.60
     اولین
    0.60
     каждом
    0.59
    this
    0.56
    était
    0.55
    Activations Density 0.016%

    No Known Activations