© 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. 15794
    Prev
    Next
    INDEX
    Explanations

    lead generation, democracy, legal, medical, gender identity

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

    This neuron responds to unusually uncommon or low‐frequency tokens (i.e. words that are rare in the text).

    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
     Kotlin
    0.84
     acrylate
    0.82
    崽
    0.76
     pona
    0.74
     electroweak
    0.74
     다니
    0.74
     calves
    0.73
     Цуки
    0.73
     Tsai
    0.72
     Summon
    0.72
    POSITIVE LOGITS
    unfortunately
    0.88
    université
    0.81
    żd
    0.81
    Unfortunately
    0.81
    university
    0.80
    此之外
    0.80
    ن
    0.80
    è
    0.79
    j
    0.79
    UNESCO
    0.77
    Activations Density 0.000%

    No Known Activations