© 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. 4914
    Prev
    Next
    INDEX
    Explanations

    past and future contexts

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

    The neuron fires strongly on tokens with very large vocabulary IDs (i.e. rare/uncommon tokens), regardless of their actual lexical content.

    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
    vykor
    0.89
    㕮
    0.86
    ніка
    0.86
    avasena
    0.83
    czną
    0.83
     యొక్క
    0.82
    oggetto
    0.80
     Использу
    0.80
    建筑物
    0.80
    πάρχ
    0.80
    POSITIVE LOGITS
    ே
    0.89
     diabet
    0.89
     morgen
    0.86
     ny
    0.80
     apoi
    0.80
     mañana
    0.80
    :
    0.79
     domani
    0.79
     a
    0.78
    ia
    0.78
    Activations Density 0.000%

    No Known Activations