© 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. 31-GEMMASCOPE-2-RES-16K
    7. 12805
    Prev
    Next
    INDEX
    Explanations

    dinosaur, dragon, or reptilian creature

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

    The neuron strongly activates on tokens containing “dino” (e.g. dinosaur, dinobot, Jurassic), effectively spotting references to dinosaurs.

    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_31_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
     Laplacian
    0.70
     काली
    0.64
     अभिन्न
    0.60
    impurity
    0.56
     categorical
    0.56
     Colombe
    0.55
     ಒ
    0.54
     कुमारी
    0.54
    promotion
    0.53
     Tolst
    0.53
    POSITIVE LOGITS
     dinosaur
    1.90
     Dinosaur
    1.78
     dino
    1.71
    🦖
    1.68
     dinosaurs
    1.65
     Jurassic
    1.60
    Jurassic
    1.56
    🦕
    1.48
     Dino
    1.31
    inosaur
    1.30
    Activations Density 0.081%

    No Known Activations