© 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. 7572
    Prev
    Next
    INDEX
    Explanations

    json formatting

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

    The neuron fires strongly on capitalized tokens—especially proper names and acronyms—i.e. it detects proper nouns and uppercase terms.

    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
    ডিও
    0.77
    ialmente
    0.71
    りで
    0.65
    вов
    0.64
    antara
    0.62
     antih
    0.62
    egra
    0.62
    夊
    0.62
    ству
    0.61
    Integrante
    0.61
    POSITIVE LOGITS
    webtoken
    0.98
    rpc
    0.77
    stringify
    0.76
     toJson
    0.73
    RPC
    0.70
     json
    0.68
    parse
    0.66
     Bapt
    0.66
     parse
    0.66
    minify
    0.64
    Activations Density 0.054%

    No Known Activations