© 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. 7929
    Prev
    Next
    INDEX
    Explanations

    numbers with operators

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

    The neuron is basically a “content‐word” detector—it lights up on informational tokens (nouns, verbs, names, numbers) and stays off on 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
    America
    0.66
    0
    0.63
    
    0.62
    Washington
    0.61
    ^
    0.61
    incinnati
    0.60
    pubs
    0.60
    http
    0.59
     either
    0.59
     Harrisburg
    0.59
    POSITIVE LOGITS
     🥰
    1.07
     ২০২৩
    1.05
    🥰
    1.04
     ২০২২
    1.02
     qPCR
    1.02
    🤭
    1.00
    🤎
    0.99
     Tiktok
    0.97
     eSIM
    0.96
     decarbon
    0.95
    Activations Density 0.191%

    No Known Activations