© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 8295
    Prev
    Next
    INDEX
    Explanations

    centuries and percentages

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

    The neuron fires on ordinal number tokens (e.g., “29th,” “21st,” “Fourth,” “Sixth”), i.e. numeric or written ordinals.

    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
     दर्जन
    0.62
     dozen
    0.61
     fewer
    0.61
    씩
    0.61
     thirds
    0.60
    ずつ
    0.58
     hundred
    0.56
     каждая
    0.56
     thousand
    0.55
    dozen
    0.55
    POSITIVE LOGITS
     init
    0.43
     esp
    0.43
    system
    0.41
    Jill
    0.41
    getContent
    0.40
    Finance
    0.40
    title
    0.39
    <?
    0.39
     unfold
    0.39
    Navigation
    0.38
    Activations Density 0.231%

    No Known Activations