© 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. 24-GEMMASCOPE-2-RES-16K
    7. 14597
    Prev
    Next
    INDEX
    Explanations

    The neuron spikes on unusually rare or low‐frequency tokens—i.e. it flags words/subwords that the model sees very infrequently.

    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
     producen
    0.94
     baratos
    0.93
     proporcionan
    0.89
    ﺖ
    0.88
     griega
    0.85
     kov
    0.84
    ﻞ
    0.84
     representan
    0.83
    ApiClient
    0.80
     umfasst
    0.80
    POSITIVE LOGITS
    {
    0.86
     Bacon
    0.72
     İstanbul
    0.71
    [
    0.70
     {
    0.68
    n
    0.67
    İstanbul
    0.66
    ias
    0.66
     Holes
    0.66
     .\
    0.64
    Activations Density 0.000%

    No Known Activations