© 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. 151
    Prev
    Next
    INDEX
    Explanations

    method returns or see below

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

    The neuron is highly sensitive to uncommon or unusually written tokens—especially proper names, acronyms, or other out‐of‐ordinary words—firing strongly only on rare/proper‐noun‐style tokens.

    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
     下
    0.84
     這
    0.81
     车
    0.75
     桌
    0.74
     vše
    0.74
     läbi
    0.73
     휴대
    0.72
     주
    0.72
     만들어
    0.72
     पटाखे
    0.72
    POSITIVE LOGITS
     apprehend
    0.87
     guardianship
    0.79
     priors
    0.77
     desist
    0.72
    קת
    0.72
    normalize
    0.71
    plify
    0.71
     nominees
    0.70
    তাহাদের
    0.70
     apprehended
    0.70
    Activations Density 0.000%

    No Known Activations