© 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. 11501
    Prev
    Next
    INDEX
    Explanations

    training, formation, designed, trigger

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

    This neuron fires strongly on rare or unusual subword tokens—especially numerical strings and obscure proper-noun fragments.

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

    tokens that are part of Gaelic or non-English language words.

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

    mentions of specific languages or linguistic adjectives (especially language names ending in -ic).

    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
     Targets
    0.86
    >
    0.85
    jit
    0.80
     targets
    0.78
     adjourned
    0.78
     translateX
    0.76
     getters
    0.76
     రా
    0.75
     opponents
    0.74
     prostatic
    0.74
    POSITIVE LOGITS
    льне
    0.81
     снаб
    0.75
     بخشی
    0.73
    すすめ
    0.71
    不够
    0.70
    واره
    0.69
     тә
    0.67
    供应
    0.66
     ramach
    0.66
     disciplina
    0.66
    Activations Density 0.000%

    No Known Activations