© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11719
    Prev
    Next
    INDEX
    Explanations

    broken down

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

    This neuron responds to occurrences of the word “broken” (in any capitalization or morphological form).

    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
     breaking
    0.80
     brak
    0.72
     Breaking
    0.67
    breaking
    0.64
    剎
    0.62
    break
    0.61
     解
    0.60
     break
    0.60
     BREAK
    0.60
     tearing
    0.58
    POSITIVE LOGITS
    StreetMap
    0.61
    📐
    0.58
     simpat
    0.58
     magnificence
    0.58
    Dreams
    0.57
    📏
    0.56
     Dreams
    0.56
    🌏
    0.55
    applique
    0.54
    🌍
    0.54
    Activations Density 0.017%

    No Known Activations