© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 4107
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on uppercase tokens marking news‐style datelines or place names (e.g. city or state names at the start of a paragraph).

    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
     essendo
    0.93
     mieć
    0.87
     BSc
    0.87
     być
    0.85
     juga
    0.84
     superficies
    0.83
     essere
    0.83
     く
    0.82
    🥴
    0.82
     calcareous
    0.81
    POSITIVE LOGITS
     congressional
    1.35
     Washington
    1.23
     ACLU
    1.18
     Obama
    1.17
     bipartisan
    1.14
     WASHINGTON
    1.13
     lobbying
    1.13
     Congressional
    1.11
     political
    1.09
    Obama
    1.09
    Activations Density 0.057%

    No Known Activations