© 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. 24-GEMMASCOPE-2-RES-16K
    4. 4455
    Prev
    Next
    INDEX
    Explanations

    potential buyers

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

    The neuron activates primarily on numeric tokens (especially decimal‐formatted numbers).

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

    the word "on" when it appears in contexts related to locations, times, or surfaces where actions take place.

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

    high-frequency function words, especially short prepositions, possessive markers, and other grammatical connectors.

    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
    ಯ
    0.80
    CLAIM
    0.79
     Himself
    0.78
     Фо
    0.77
     Его
    0.75
    IONE
    0.73
    Mein
    0.73
    CCc
    0.73
     Bloss
    0.73
     задума
    0.72
    POSITIVE LOGITS
    ść
    0.82
    いた
    0.79
    itation
    0.77
    ning
    0.76
    تالي
    0.75
    所谓的
    0.74
    ました
    0.74
    க்கூடாது
    0.74
    ment
    0.73
    굶
    0.73
    Activations Density 0.000%

    No Known Activations