© 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. 7162
    Prev
    Next
    INDEX
    Explanations

    code `id` or `find`

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

    The neuron fires strongly for tokens appearing very late in a long input sequence—that is, it detects words at high absolute token‐position indices.

    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
     ସ
    1.00
     ଦ
    0.96
     vivió
    0.96
     volver
    0.96
     เร
    0.92
     oluşan
    0.92
     ahorrar
    0.92
     ଗ
    0.92
     veröffentlicht
    0.91
     yılı
    0.90
    POSITIVE LOGITS
    ENO
    0.74
    SF
    0.73
    бари
    0.70
    MEN
    0.69
    NES
    0.68
    ELY
    0.66
    ELINE
    0.65
    RED
    0.65
     гости
    0.65
    elines
    0.64
    Activations Density 0.001%

    No Known Activations