© 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. 3742
    Prev
    Next
    INDEX
    Explanations

    foreign and religious figures

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

    The neuron sharply activates on relatively rare or content‐heavy tokens—often descriptive adjectives, proper nouns, or hyphenated/compound words—marking semantically rich or unusual words.

    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
    ből
    0.99
    y
    0.93
    DVRIP
    0.92
    م
    0.91
     Badge
    0.88
    n
    0.88
    ievement
    0.87
    י
    0.87
    ي
    0.85
    ต์
    0.84
    POSITIVE LOGITS
    Rhe
    0.71
    Krishna
    0.71
    Ma
    0.66
    Schwe
    0.66
    izons
    0.65
     swat
    0.65
    хар
    0.64
    Emma
    0.64
    omag
    0.63
    Idaho
    0.62
    Activations Density 0.001%

    No Known Activations