© 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. 2247
    Prev
    Next
    INDEX
    Explanations

    calculation symbols

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

    This neuron is essentially silent on ordinary text and only “lights up” on isolated out-of-vocabulary or proper‐name tokens—e.g. “Youtube” (and similarly rare capitalized names)—so it appears to be detecting unusual or external‐reference 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
    uradaki
    0.78
    modation
    0.77
     دیا۔
    0.75
    連携
    0.71
     җиңү
    0.71
    鐙
    0.70
     ይችላል
    0.70
    jezd
    0.70
    reements
    0.70
    ो
    0.70
    POSITIVE LOGITS
    De
    0.90
     Aly
    0.86
     Zach
    0.84
     Dj
    0.84
     Al
    0.83
     Arik
    0.81
    COR
    0.79
     De
    0.79
     noy
    0.79
     thor
    0.78
    Activations Density 0.000%

    No Known Activations